Piotr Mitros
Portrait of Piotr Mitros
piotr@mitros.org

Piotr Mitros is an innovator and expert in the field of education. He has been a co-founder or key early employee at three organizations, all of which have crossed the $100 million mark. Mitros is a frequent keynote speaker, panelist, and consultant on: disruption in education, assessment, learning analytics, educational datamining, open educational resources, learning-at-scale, open education, use of evidence in education, ICT for development, and crowdsourcing in education. When possible, Mitros conducts such presentations in an interactive, group-learning format.

When not traveling, Mitros is focused on a new not-for-profit targeting access to justice, civics, and open education. He is a candidate for the Cambridge School Committee, where he hopes to use his experience in education research and in cross-cultural education to improve the Cambridge School District. He also enjoys consulting when particularly technically interesting or socially-beneficial projects come up.

Mitros is a free software author, perhaps best known as the creator of Open edX, an educational platform which has around 300 contributors, and 200 deploys, including ones from the Saudi Ministry of Labor, the Ministries of Education in France, and China, and the Queen Rania Foundation (the not-for-profit of the Queen of Jordan), Stanford, the World Economic Forum, as well as edx.org. As of this writing, it powers around 1,000 full, pure-online courses, has around 10 million users, and forms the backbone of a new research ecosystem. Designed from the ground up for educational data collection, randomized control trails, and experimental pedagogies, at the most recent Learning@Scale conference, all but one of the best paper nominees (including the winner) were based on Open edX. Mitros proposed the concept of MITx/edX to MIT in the fall of 2011, and served as the Chief Scientist of edX, the MIT/Harvard education initiative, for over a half-decade.

Mitros has a long history developing educational models, especially in diverse cultures. The concepts behind the MITx project were inspired by Mitros' experiences teaching in China, working in India, facilitating educational technology projects in Nigeria, and developing experimental educational formats at MIT. His observations of university systems around the world inspired Mitros to find innovative ways to dramatically increase both the quality of and access to education. He has pioneered several models by which groups of students and educators can work together to construct courses with improved learning outcomes, pioneered many ways to use technology to improve learning outcomes in classrooms for diverse populations, and did extensive data analysis on one of the largest psychometric datasets in the world. His pedagogies and technologies have enabled improved learning outcomes across an incredible diverse range of settings, backgrounds, and cultures.

Mitros briefly served as Director for Education at what was at the time Know Labs and is now Udacity. Udacity was valued at over $1 billion at its last investment round. Mitros developed the analog electronics for a new medical imaging modality for Rhythmia Medical. The resulting device (which also includes novel numerical algorithms, data visualization, mechanical design, etc. developed by other members of the Rhythmia team) can remotely image the surface of the heart, shaving hours off a particular type of atrial surgery, and dramatically increasing its accuracy. Rhythmia Medical sold for around $200 million and the Rhythmia catheter and patient interface unit have been used on thousands of patients saving many lives. As part of his work at Rhythmia, Mitros designed what is still most likely the highest-performance ECG circuit in the world.

Mitros is author of many peer-reviewed scientific publications, and holds a B.S. degrees in Math and Electrical Engineering, a Masters of Engineering in EECS, and a Ph.D. in EECS, all from MIT, with which he has been continuously affiliated in some way since he was a junior in high school.

Hobbies and interests

How people learn, anthropology, analog circuit design, digital circuit design, computer science, control systems, theoretical mathematics, transformational leadership, disruptive innovation. Areas of some background: physics (classical mechanics, waves, thermodynamics, and introductory quantum), signal processing, machining, wood-working, welding, soldering, sewing, bicycling, martial arts, Chinese (intermediate), Arabic (extreme beginner, despite countless hours invested).

Piotr Mitros: Scholarly Works
Curriculum vitæ. roughly translated, means course of my life. That suggests, if you'd like to know my life experience — what it really means to be me — there are few better places to look than this automatically generated page listing my talks, publications, and research interests.

Conference and Journal Publications

Other Publications

Presentations

Service, Awards, Personal

  • 2016 UM/ARL GIFT Advisory Board
  • 2016 Chair industry track at Educational Datamining
  • 2016 Blended track chair/coordinator at Learning@Scale
  • 2015 Williams, et. al. “Connecting Collaborative \& Crowd Work with Online Education” CSCW workshop co-organizer.
  • 2015 Co-chair industry track at Educational Datamining
  • 2006 Total Fellowship
(The above is semi-automatically generated from my CV)

Selected Areas of Research

Improving Education with Economies of Scale I designed the Open edX platform and pedagogy to allow efficient application of evidence-based techniques in teaching-and-learning in blended classrooms. We've had tremendous results. In a randomized control trial at SJSU, we were able to bring course completion rates from 59% to over 91%, and later, we doubled the amount of material taught. In pure on-line settings, we were able to deliver learning gains between those of traditional classrooms and blended active learning classrooms. Community college students have been able to learn the MIT introductory computer science curriculum. We've had many similar results in less formal evaluations.

Crowdsourcing in Education I've had a number of success in finding ways to bring communities of people together to create educational materials. The core difference from traditional OER approaches is that OERs try to compile curricula from disprate resources. Instead, I bring communities together around a common curriculum. Probably the most significant result is that students can contribute very hints more effectively than in commercial intelligent tutoring systems where such content costs millions to make. Probably the most important (and underappreciate) result is that students can create high-quality assessments, if given appropriate scaffolding. Probably the most important area of continued research is bringing together communities including instructors, researchers, and industry practitioners around such courses.

Circuit Network Theory Recently, I put out a very nice paper showing we can have filters with much lower error than the traditional Butterworth/Chebyshev/Bessel/etc. designs. Prior to that, my Ph.D thesis solved a number of previously unsolved problems. It described a methodology for analog circuit design that used bidirectional flow of information. This methodology allows for the construction of complex, interconnected systems, guaranteeing their stability with what is mostly a local stability criterion. I applied it to three problems. First, I showed how it could be used to solve systems of equations. Second, I showed how it could be used to build models of systems by observation. Finally, I showed how it could be used to linearize a nonlinearity. The linearization is probably the most compelling application. The circuit is fairly simple: A linearizer that can model a nonlinearity as a quadratic Taylor approximation, and predistort to compensate for it, can be built with under a dozen major blocks (Gilbert cells or op-amps). It is independent of feedback linearization, so it can be used inside of a feedback loop without significantly affecting feedback linearization. More complex linearizers are also robust to a class of device failures. Ironically, because many such problems were considered closed, there wasn't much of a community to read my thesis.

Piotr Mitros: Code

Code on Github

I am perhaps best known as the creator of Open edX, a widely used platform for education. Other assorted free software I've been involved with includes (automatically generated with github's API, but a little more organized and readable):

Projects

Utilities

Libraries

edX and Education

AnimationXBlock
XBlock for embedding something like an animation. You can browse through Images+text with a slider.
AudioXBlock
A very simple audio player for edX
concept-tag-server
A server for storing concept tags, based on django-simple-wiki. There is a corresponding XBlock.
ConceptXBlock
An XBlock for adding concept tags. There is an associated external service in a different repo.
DisqusXBlock
An XBlock for embedding Disqus discussions, to use instead of edX discussions.
DoneXBlock
An XBlock for students to mark they've finished something.
edX-Insider
OLX for a short course on online pedagogy
edx-rest
Provides a Python API to edX REST APIs
edx-speech-tools
Tools for speech recognition and alignment written by Venkatesh Sivaraman, co-supervised by myself and Juho Kim
edxml-tools
Tools for manipulating edXML/OLX, including converting it to human-friendly format, making podcasts, etc.
FeedbackXBlock
An XBlock to allow students to provide feedback on course materials
ImageXBlock
An XBlock for displaying a single image
LectureScapeBlock
XBlock for LectureScape, a video player driven by data and statistics, written by Juho Kim, and converted to an XBlock by Peter Githaiga, under our co-supervision
pres2edxml
Convert a PowerPoint/OpenOffice/PDF presentation to OLX/edXML for embedding in an edX course.
ProfileXBlock
Prototype XBlock for showing a course-specific student profile
RecommenderXBlock
An XBlock to recommend resources to other students, written by Daniel Li, under my supervision
SelfCheckXBlock
OLI-style check yourself XBlock. Stanford used this as a prototype to develop theirs.
x-analytics-scripts
Assorted analytics scripts for edX
XBadger
An XBlock for navigating courses based on objectives and badges
xblock-future
Proposed future functionality for xblocks. I toss things in here which are hack-arounds for missing XBlocks functionality.
xblock-submit-and-compare
A self assessment XBlock for the edX Platform which allows students to submit and self-grade freeform text.

Mostly Older and Obsolete

Other Projects

I've also:

  • Built scientific software to process data from oil wells for Landmark Graphics Corporation (as a high school student)
  • Built the software for a web startup for Boston Consulting Group (to cover tuition)
  • Built a bunch of numerical algorithms for Rhythmia Medical (now Boston Scientific) primarily around things like compensation for signal issues and calibration of hardware.
  • Built software to automatically optimize (and to some extent design) analog circuits based on gradient descent (for component values) and simulated annealing (for both component values and circuit topologies), primarily to assist with design of 3D integrated circuits. My master's thesis talks more about it.
  • Built an early web application in the late nineties which allowed an elderly community to publish a newspaper easily and with good workflow (as a high school student, at the MIT Media Lab)
  • Made significant contributions to a project to allow students in developing nations to remotely control laboratory equipment (MIT iLabs project)
  • Developed plug-ins for Internet Explorer (ATL / COM / MFC / Visual C++) and high-performance Apache extensions in C
  • Done a fair amount of low-level development, much of it in assembly, some of it on oddball platforms (e.g. handcoding assembly for a 4x4 grid processor)
  • And, when I was a freshman, I spent a large amount of time designing algorithms to localized and separated sound sources using a microphone array, relying on approaches such as modifications of aperture synthesis, as well as phased array approaches. It never worked especially well.
Piotr Mitros: Creations

For most of my life, makers were called engineers. This was a little bit ridiculous, seeing as few of us worked on train engines any more. Thing-makers wore jeans and button-up shirts with pocket protectors, carried large calculators, and worked in laboratories or workshops. A few years back, we donned turtlenecks, and rebranded outselves as makers, and our workshops, as makerspaces.

Amazingly, it worked! It did wonders for our image. So, like a Gnome in a workshop, I am now a maker of things. A thing-maker. This page lists my things.

Rhythmia Analog Front End

Rhythmia Signal Acquisition Unit

I designed almost all of the analog electronics for a new medical imaging modality — from blank slate all the way through human trials. It allows doctors to make physical and electrical maps of the heart in a fraction of the time and with much higher accuracy than existing techniques. The system took measurements on a large number of electrodes on a catheter inside the heart (nearly 100), and extrapolated the signals on the heart's surface. The system continued to work during cardiac ablation with no visible signal degradation (maintaining sub-microvolt RMS noise floors in the cardiac band in the presence of 300V peak-to-peak interferer at 500kHz). Some patients are defibrillated many times during the procedure, so we had to survive a 5kV spike for about 3ms. The catheter went into the heart through an incision in the leg, and was threaded up into the heart. All the signals were cross-coupled, and we had to make sense of all of that. It's been used in a few thousand surgical procedures at this point, and has saved a fair number of lives. The system included what was likely the world's highest-performance surface ECG as well, which is surprisingly easy to do (design of ECGs has advanced surprisingly little in the past few decades). Minor footnote: If you look on Boston Scientific's web page, they'll rate the noise a good deal higher. This is mostly because of the funky way in which medical devices rate noise (which is based on 98th percentile peaks, rather than RMS).

Bicycle with Custom Handlebars

Hand-made bicycle with custom-welded handlebars

I made myself a bicycle. Or specifically, I bought parts, and assembled it myself. It's got an internally geared hub, a Shimano Biopace eccentric crankset, tandem wheels (lacing up and truing wheels is a Zen kind of relaxing), front and rear rack, carries an infant, and is generally the toughest, burliest bicycle you'll see. For better ergonomics, I also welded the handlebars myself — I held out my hands, and put rods of metal where my hands most comfortably went. I'm pretty proud of how nice the TiG welds came out. The inside of the frame is lined with linseed oil, as are most of the nuts and bolts. The nice thing about that, and choice of other parts, is I can leave it out in the winter, in deep snow, in the rain, and all sorts of other conditions, and it does fine. I can ride on almost flat tires. I don't need to worry about it. It's spent winters outside, and it's done just fine (although recently, I did sew a nice cover for it).

Playset

Home-made playset with a punching bag, music band, and three-person swing

My (2.5-3.5 year old) son and I made a playset! And he really could help! He did a lot of the painting. He'd place a square, I'd adjust it and draw a line, he'd centerpunch a hole, I'd drill it, he'd put a screw in, and I'd screw it. At 2.5, he could center punch a whole with RMS error of under a millimeter. Little guy's got a good attention span. It's held together by structural screws. I actually didn't know those existed before — I found out when looking for good fasteners to use which wouldn't corrode in pressure treated wood. Most screws sheer very easily (and just pull wood together — for sheer, we use nails or fancy joinery). In contrast, good structural screws will hold around a thousand pounds each. I aim for about 10x weight for dynamic vs. static loading, and that's pretty easy with structural screws. I have large diameter bolts in just a few places where there may be more extreme stresses. It's got a swing where we can read books together, a punching bag (he really wanted after using one regularly at the gym with his mom), a few climbing structures, and a band set (including a wooden xylophone).

Wooden Bed with Adjustable Legs and Head

Adjustable bed. Shown with one end raised. It's not obvious from the photo, but it's actually the legs which are raised in the photo, not the head. Otherwise, I would have had to rotate the bed 180 degrees for the snapshot -- a major hassle

This is my bed. I made two of them, but an ex has the other one. It's super-sturdy, and the head and feet go up and down. It also has outlets, so you don't need to run wires all over the place to power your laptop, cell phone, iPad, etc. in bed. The frame is 4x4, with 1x4 slats, and a second folding frame on top of that. It's a bit over-engineered, but that way, (1) it doesn't shake or squeak at night (2) it has plenty of room underneath it to vacuum (or for storage) since the legs don't need additional cross-braces.

Optical Encoder

Motor with an optical encoder attached. The optical encoder consists of a custom printed circuit board, a laser-cut acrylic wheel, and a laser printed piece of paper attached to the wheel

I redesigned the reflective optical encoders for MASLab. They were fully differential (they pattern had two strips for each phase instead of one — when one was black, the other was white, and the circuit compared the two). As a result, they did not rely on any absolute threshold, and could be used in a variety of lighting conditions, and with the reflective surface anywhere from several millimeters to several inches from the sensor. Perhaps ironically, this is one of the highest-impact projects I've worked on. It was a very quick project, but for a long time, it was one of the very few (and the best) DIY optical encoder designs on-line, and was quite widely adopted, in all sorts of oddball places (like Hollywood productions). I've even been asked to review IEEE journal publication on optical encoders as a result of this — a field where I'm clearly no expert. The project included a custom printed circuit board, laser-cut wheel, and pattern.

Autonomous Robot

A robot which looks suspiciously like a bunny

I was on a four-person team for the Mobile Autonomous System Laboratory at MIT. It is a month-long contest to develop an autonomous robot capable of completing a task that varies from year to year. Our year, the goal was to bring back a series of beacons in an unknown environment. Our robot was the only one to successfully bring back a target beacon.

Metal+MDF Sewing Machine Table

sewing machine table. Visible is an aluminum top, a leg (which was welded), and a complex structure for raising/lowering a sewing machine

I made a sewing machine table for my sewing machine. The table top is MDF covered with sheet aluminum. The legs are oxy-acetylene welded steel. There's also a structure for raising/lowering the sewing machine, so it sits flush with the table.

Welded Monitor Stand

A monitor stand with five monitors, angled in for easy viewing

I made my monitor stand by welding a bunch of steel pipes together with a $100 Harbor Freight flux core welder. Welds are ugly, by a welder's aesthetic (which also means they wouldn't hold up hundreds of pounds), but they gets the job done, and they're not aesthetically displeasing unless you know what a technically proper weld should look like. There's a 4k monitors in the middle, surrounded by 1080p monitors, two portrait and two landscape. It has a little bit of support for clean cable routing too.

Speakers

A very ordinary-looking pair of speakers. Black

After I took an acoustics class from Bose (the same Bose who ran the speaker company also taught at MIT), I wanted to apply what I learned, so I made a pair of speakers. It's a pretty standard design, except for the external crossover (seen on top). Using an external crossover lets me (in abstract) put the crossover before amplification, using lower voltage (and more linear) components in the crossover.

Classical Breadboarded DC-DC Converter

A pair of cutting boards for bread, with a circuit schematic laser-etched onto them, and an actual circuit soldered on top

Ever wondered where the term breadboard came from? As it turns out, people used to prototype electronics by sticking brass tacks into an actual breadboard and welding to them. This is a demo of a boost converter made using a modern take on the same technique. It's for students, so I laser etched the circuit schematic into the breadboard below the components. After I made one, a student made another, with a less subtle etching.

Sewing Iron Bag

A bag for an iron, decorated with a frilly finish, and a CNC-cut robot (colored with permanent markers)

I was given an iron for keeping my clothing creased. I made this little padded bag for my iron to keep it safe. It's a simple padded bag. I've made many much like it for my other tools too. The little robot was CNC-cut.

Turbidistat

I made a turbidistat. What's a turbidistat? It's a device which measures turbidity to estimate bacterial concentration in a fluid. If the bacteria get too dense, it pours out some of the bacteria, and pours in some medium for new bacteria.

Signal Acquisition Package for African Universities

An unimpressive-looking printed circuit board

I once made a board very similar to an Arduino, sometime before the Arduino existed. It has a little more signal functionality in terms of ADC/DAC. It was designed to be used as a (low speed) power supply, function generator, and oscilloscope at a university in Africa.

Fabric Games

Pieces of fabric cut and stitched together. One has a set of holes to wind a twine through and such a twine. The other has a 4x4 grid for round pieces, and a set of pieces to the side

I have a fabric cutter at home. A family member wanted to make a few puzzles out of wood. I made the same puzzles out of fabric. One is a connect four style game, where pieces four properties (hollow/full, color, shape, whole/half). In the other, you try to find a path which passes through each node once. I make a lot of things on my fabric cutter. What's need is a project like this takes something like an hour start to finish. What's not so neat is that all the fabric cutters have locked-down proprietary file formats and interfaces.

Bookcase

I made a rather nice bookcase. I was just learning to woodwork, so it took a rather long time. All of the screw holes have plugs (wood of the same color above the screw head), so they're almost invisible. The shelves adjust in height (except for the middle one). The back is in-set. And the joinery is all rather fancy.

Desk

I made a desk out of an old workbench tabletop and some poplar legs. I had to replane the tabletop quite a bit. The top is also routed so it looks like it's floating.

Sheet Metal Speaker Stand

A cheap JVC speaker on top of a nice stand made of sheet aluminum, cut and folded to hold it, and painted black to match

I made some speaker stands out of sheet metal. They're for a pair of old speakers someone gave me when I was in college.

Quiet Computer

In the late nineties, I was a leading authority on making near-silent computers. At the time, I slept in a room with a computer. Computers were noise; it kept me up at night. I found ways to make it much quieter (which involved things like retrofitting industrial power supplies, finding/retrofitting convection heat sinks, using mobile CPUs with desktop motherboards, finding ways to isolate hard drive noise, etc.). As I recall, I only broke something once; I had my hard drive spin down when idle, and it turned out nineties-era hard drives would fail after a relatively small number of start/stop cycles. I also shocked myself a couple of times working with open frame power supplies before I really knew what I was doing (never causing any damage; 120VAC is pretty safe if it doesn't pass through the heart or perhaps the brain). There were 2-3 major web sites on the topic, and I ran one. Later, mainstream manufacturers started paying attention to acoustics, and now there are many off-the-shelf silent or near-silent computers. If you want one, you can get something off-the-shelf from Pudget Systems or similar.

RFID Hack

I made a little device which emulated an MIT ID (specifically, Jerry Sussman's, with his permission), so I could have access to MIT after I graduate. This was based on a design by Keith Winstein, but somewhat enhanced (mostly a nicer form factor through the use of a planar antenna I designed, rather than a large coil as Keith used).

VCO Design

I spent the summer of 2004 working for Texas Instruments in India on VCO design. I'll post more information about what I did there at some point (I need to sort out what's NDAed).

3d Video Game on an FPGA

A photo of a CRT with a 3d racing game, consisting of a ball with a shadow on a track. The track has two levels, of which the bottom is texture mapped, and the top is transparent. There is an animated background behind all of this, but the animation is not obvious from a static photo

Balkentrol is a 3d racing video game I designed and built. It's implemented entirely in hardware (no microprocessor or computer of any sort; just an FPGA (programmed in VHDL), some DACs, op amps, solder, wire wrap, resistors, etc.). I implemented this while I was in 6.111, and turned it into my final project. Later, I was joined by a partner who built a fancy sound system for the game. To do this, I had to overclock the FPGA to run at the VGA dot clock, so the VHDL had to be pretty highly optimized. This project was nominated by the professors for an award as the top MIT EECS lab project that semester.

Laser Cutter Art

A slice of a tree trunk with two people etched into it

I used to do a fair bit of artwork using a laser cutter. The above is a picture of my father with his wife cut into a slice of a tree.

Instrumentation Photosensor

Working for Talking Lights, I designed and built an instrumentation photodetector, where the noise was dominated by the shot noise in the photodetector diode (caused by discretization from individual photons striking the sensor). It had several MHz bandwidth. It was designed to measure very small signals on top of a large constant light source.

Assorted Nifty Circuits: Power Converter

A simple analog circuit schematic

Power converter In Roberge's Advanced Circuit Techniques course, I designed a fairly novel power converter. It added an idle state to move a zero in the right half plane into the left half plane, allowing for very high performance. The initial design, aside from this change, was over-ambitious in a few other ways (self powering, etc.), and as a result, I ran out of time, so I ended up implementing a different design at around the deadline, and finishing the original design a bit later. A design identical to mine was later independently published by IEEE Transactions on Power Electronics. I wish I'd thought to publish it — it Would have been a nice to have a very early career paper!

Assorted Nifty Circuits: Line Driver

A complex analog circuit schematic

Line driver In Roberge's Advanced Circuit Techniques, I designed a high speed line driver. It was a one-shot design (components were chosen to match each other, and it would not work with randomly chosen parts), but it had a small signal bandwidth of 60-70MHz, on a protoboard (about 5x the speed of the fastest project people were aware of from current and past semesters, and if I recall correctly, and 30x faster than spec). Of course, for large signals, it slew-rate limited as well (although the slew rate was still very, very significantly above spec).

Less Nifty Circuits: Other 6.331 projects

6.331 is MIT most intense, advanced analog course for graduate students. These are my other projects. They met specs, but unlike the above, were fairly uninteresting in architecture:

  • Sample and hold design — had a trivial DC biasing bug, but otherwise was fine. Could meet both sets of specs with the change of a capacitor (we were required to meet either a high-speed or high-precision spec).
  • Digital to analog converter design — completely boring design based on a current-switching architecture. Had two mistakes in the report (not the design). I included the wrong version of section 4, which had a spurious 2000pf cap, and I accidentally wrote "potentiometer" in one place where I meant "resistor"
  • Analog to digital converter design — Flash DAC followed by a bunch of 1 bit DACs. Don't recall why it made sense to reverse the order and put the flash DAC at the beginning rather than at the end.
  • Phase locked loop lab — Completely cookbook lab.

Carry-lookahead adder

In 6.374, for the final project, my team designed an adder based on a scheme I came up with that computed carry for 3 stages at a time, using what was effectively a DAC and comparator (implemented as a sense amplifier, with multiple input transistors, with scaled W/L ratios for the 3 inputs). Good idea? Bad idea? Hard to say, but it was certainly an interesting idea. The class was fortunately taught by Tom Knight, who valued creative and interesting approaches.

Microcurrent sources

In 6.973 (subthreshold VLSI), I designed a series of microcurrent sources. They worked by generating a single precision current with a bandgap reference and an off-chip resistor. This current was converted into an oscillation frequency (based on a voltage reference and a capacitor value). The frequency was digitally clocked down, and converted back to current values (dependent on the new oscillation frequency, as well as a matched capacitor, and the same reference voltage). At the time, I had a fairly weak background in control systems, and as a result, this isn't one of the projects I am more proud of (this solution, in retrospect, was substantially underanalyzed, and so likely quite poorly optimized). On the bright side, the paper was fairly well written.

Audio Visualization Hardware

When my sister came up to visit, we hacked together an analog audio visualization circuit. We built a beat detector, and drove the deflection coils of an old CRT from a Mac Plus to draw a circle that got bigger and smaller in response to the beat.

Function Generator Improvement

The FG500 is an ultra-low-cost function generator. It's also battery-powered, which is super-useful, since it has no ground reference. Sadly, it has a nasty DC offset. I modified it to get rid of the DC offet.

Minor projects

I've got a billion other things I've built, for which I don't have photos or documentation. On the electronics front, I made a nice audio power amplifier once (pretty novel control loop). Bitstream DAC. AM receivers. Misc. DC-DC power converters, linear power supplies, regulators, etc. Several operational amplifiers. Temperature probe. Digital FIR filter. FPGA programmer. Simple switch-cap filter. Etc. On the mechanicals, I made much of the furniture in my various homes over the years, and many toys for my son. At the time, I didn't think to take photos...

Edge Detector on an FPGA

An MIT professor wanted to move machine vision algorithms from a computer to an FPGA to save power in a flying quadcopter. As a weekend hack, I build a system which takes video from a camera, does edge detection, and outputs to a little LCD TV. He offered me a Research Scientist position doing robotics at MIT. I almost took it, but ended up starting edX instead. If I'd taken that, I might now be a highly paid MIT-trained roboticist instead of a not-for-profit educator!

Garden

A little broccoli head growing in my garden.

I have a garden! Successes this year include tomatoes, radishes, mint, chives, lettuce, strawberries, mulberries, black raspberries, and peas.

Graphic Designs

I spent four years on my high school's newsmagazine. I mostly did graphic design and photoediting. Can you believe I used to be a pretty good designer? Hard to believe today, looking at my web page today....

Power Controller

MIT Solar Car I was drafted for a fast-paced 2 weeks to rewrite the control code for the microcontroller controlling the lithium ion batteries for the MIT Solar Car team. The folks there wrote the code, having no working hardware, and left, having never tested it. As expected, it didn't work, and had a number of minor and major bugs. I debugged/rewrote it so it worked. I also helped with soldering, debugging analog electronics, etc. a bit.

Unmakes

And as a thing-maker, I'm sometimes a thing-fixer and a thing-unmaker. In the process, I discovered what goes into Sony "1080p " video, and into off-brand "120W" speakers.

Conclusion

And I'll finish up with a picture of glassworking:

A picture of Piotr Mitros in a glassblowing shop doing something with a hot piece of glass
Piotr Mitros: Cultures

Cultures are a bit of a hobby of mine. At this point, I've visited Mongolia, China, Sweden, Qatar, Israel, Mexico, France, Spain, Belgium, Netherlands, Canada, Guatamala, Greece, Scotland, Ghana, Togo, Benin, Costa Rica, Panama, Ukraine, Lithuania, Poland, India, Nigeria, and the United States. I don't travel much for tourism — I try to find ways whereby I can embed myself in the cultures, which mostly means traveling for work, on collaborations, or to visit friend/family.

At some point, perhaps, I'll take the time to write up all of these, but for now, I'll give a few highlights.

Nigeria

A photograph of a sign at Obafemi Awolowo University Ile Ife which reads Campaign Against Anti-Social Acts/The bonus and skulls destroyed by cultists belong to someone like you/Support the campaign against cultism

As a graduate student (although not related to my thesis work), I spent a few years working on MIT iLabs, a project to develop laboratories which could be teleoperated from the developing world. Before my time, this had started as a way to save MIT students the bother of having to walk all the way from their dorm rooms to MIT's labs. This may sound frivolous, but it was not. In electrical engineering, it is important to build intuition for how mathematical models relate to real-world behavior. Consequently, engineering students could work with real-world data on virtually all problem sets, rather than a few times per semester. At some point, someone noticed such labs could also be used in parts of the developing world where such equipment was simply not available. I joined the project and worked closely with Obafemi Awolowo University in Nigeria. This gave an opportunity to explore the education system there (often, in way unrelated to iLabs, such as visiting K-12 schools there), and it's actually where the idea for edX came from.

OAU was a fascinating place. Just to give a little bit of context, it's the second best school in a country of more than half of the population of the US. Nigeria has a reputation as a rough country — the State Department manages a list of around 30 of the world's most dangerous countries, many being active warzones, but Nigeria always somehow makes it there. My guidebook on Sub-Saharan Africa listed, as one of the highlights for Nigeria, "Making it out alive," and I had my share of adventures. The students cared about learning, and many of the faculty were skilled, passionate, and hard-working, but many constraints — from student-faculty ratio, to availability of equipment, to curriculum, to student's prior knowledge — meant students didn't do very well. At the time I visted, and ex-par from Europe or the US would earn about 50 times as much as an OAU graduate with a similar degree. As with much of the developing world, each time I visited, the country was visibly better than before — better economy, less corruption, and less crime.

Poland

I was born in Communist Poland, and I continue to visit Capitalist Poland every few years. This gives a perspective on the evolution of cultures. I miss many of the changes, living in Massachusetts day-to-day, since they happen so gradually. Perhaps the most interesting change in the transition to capitalist democracy was the rise of materialism. It turned out that having a multi-billion dollar advertising industry trying to convince people that their self-worth is defined by what they own has the intended effect. On the other hand, traditional Polish culture was slightly negative. For example, the Polish National Anthem begins with the words "Poland has not yet been destroyed." Probably the most popular Polish folk song is about missing Ukraine. Polish culture became much more optimistic in recent years. I have my own theories for why that is...

China

This is a quaint little pavilion over serene West Lake in Zhejiang Province. Framing and lighting make it look a little bit spectral in this photo when printed big. The small version in the page doesn't do it justice. I should either make it bigger, or replace with a photo

I taught a short 6-week computer science course at Dalian University of Technology, and later at Zhejiang University. I also made a whirlwind tour of much of the country, mostly by train, where I went as far south as spending a week in village near Xiamen (which has just installed pipes to bring in water from a nearby river), and going as far west as Urumuqi (mostly by trains, usually in the least expensive class, which gives the greatest cultural exposure).

Ukraine

For a while, I remotely managed an engineering team in Ukraine. The team built interactive tools for specific disciplines on edX. We had interactives for letting students do things like enter crystallography planes, work with chemical equations, or visualize probability distributions. According to our customers, the team was quite successful, and many of our customers were asking to pay for more of their time out of their own funds (this was in the early days — just as MIT and Harvard had put in $30 million each — and it would have been unpolitical to collect money from MIT/Harvard professors, so we did not do this, or even have mechanisms for doing so). We had morning standups every day, which became longer meetings if necessary, and I visited the Ukraine as well. Eventually, edX had a re-org, and the team went to other managers, and was later dismantled. A few of the members went on to found RacoonGang, which is now one of the main three Open edX consulting shops (together with OpenCraft and ExtensionEngine).

India

I spent three months working for Texas Instruments in Bangalore, and then made whirlwind tour through Goa, Mumbai, Rajastan, Delhi, the Himalayas, and Agra. India was my first major foray outside of Western culture, and I left my heart there. The culture is tremendously different from the US — more so than any place I've visited. It's incredibly safe and friendly. And it's easy to get by on English.

Israel

My grandmother was khassidey umot ha-olam, so I have relatives there through adoption. I spent a bit of time there, getting to know my extended family.

Mongolia

A version of 6.002x was used in Mongolia when edX was first starting. High school students at the Sant school successfully finished the course under the mentorship of Tony Kim, and it was used at the National University as well, and I wanted to see how well these worked.

And a map...

And an obligatory map showing the places I've been:

Create your own visited countries map or check out the JavaScript Charts.
Piotr Mitros: Organizations

Since graduation, I've been involved with three organizations:

edX, Chief Scientist and Co-Founder (or, Research Scientist, MIT). Accomplishments:

  • Developed one of the top learning platforms in the world. Platform is free software/open source. Platform has around 10 million users, 200 deployments, 300 contributors, 1000 courses, and is quickly becoming the predominant education research platform
  • Crafted core pedagogy, together with Sussman, Terman, and Agarwal, which has since lead to very large learning gains in rigorous evaluations, in both in-person and on-line settings.
  • Since, I've lead research and development initiatives around topics such as educational data mining, learning analytics, learnersoucing, automated essay grading, peer grading, and intelligent tutoring systems.
edX has raised over $100 million in funding.

Director of Education, Know Labs (now Udacity). I (briefly) helped get Udacity off the ground when they were still Know Labs. Udacity is currently valued at over $1 billion.

Senior Research and Development Engineer, Rhythmia Medical. My first job after graduation, I was one of their earlier employees. I developed all of their analog hardware for a new medical imaging modality from conception to human surgical trials, including highest-performance ECG circuit in the world. I also helped with some of the numerical algorithms. Rhythmia sold to Boston Scientific for around $200 million.

Pre-graduation

Prior to graduation, I worked on voltage-controlled oscillator design for Texas Instruments/India, on photodetector design for Talking Lights in Cambridge, MA. In addition, I did consulting work to BCG, helping them incubate a startup, and later, to that startup. I was saddened when that startup pivoted, sometime after I left, to become one of the largest spyware vendors in the world. Otherwise, I had a lot of oddball research jobs for assorted MIT professors.

In high school, I developed scientific software for Landmark Graphics Corporation used by petroleum engineers to help analyze data. I also had an internship in high school at the MIT Media Lab.

Piotr Mitros: Teaching Experience

In 2013, I ran a short on-line course known as Maker Physics. The course focused on on-line pedagogy and physics education research for physics instructors. These instructors, in turn, were asked to work together to apply the materials to create open, reusable digital resources for use in physics courses. This was together with David Cormier of cMOOC fame. This was one of my many attempts to crowdsource educational resources.

In 2012, I co-taught MITx 6.002x with Jerry Sussman, Chris Terman, and Anant Agarwal. This was the first MITx/Open edX MOOC, and the fourth xMOOC. It was based on active learning, constructive learning, authentic assessment, immediate feedback, and rapid remediation of misconceptions. We had 154,763 registrants, of whom 69,221 looked at the first problem set, 26,349 got at least one point on it, 13,569 looked at the midterm, and 7,157 earned the first certificate awarded by MITx. The course was also used — extremely successfully — in places from state schools to high schools in Mongolia to on-campus at MIT. 96% of students who finished (note the sample bias -- but on the upside, reporting rate was in the nineties) reported it was as good as or better as a normal residential university course. In a randomized control trial at SJSU, the course used in a blended setting brought completion rates from 59% to over 90% (and in future semesters, allowed SJSU students to master the full MIT course).

In 2011, I helped out with the Stanford AI Course. This was the first xMOOC.

In 2006, through CETI, I co-taught an abridged version of 6.001, MIT's famous introductory computer science course, at Dalian University of Technology and Zhejiang University (China).

From 2005-2007, I worked on the MIT iLabs project, where we developed on-line laboratories for use in the developing world in close collaboration with Obafemi Awolowo University in Nigeria. I went to Nigeria three times, as a result, and saw a few other parts of Africa. I also developed some very low cost hardware for use in such laboratories, in many ways similar to the modern Arduino, but with more signal acquistion/generation capacity.

In 2003, I helped Hal Abelson and Jerry Sussman make 6.002ex, a radical revamp of the electronic engineering introductory course 6.002 where students learned circuits and electronics by working on design projects in small groups under mentors from industry.

I was also involved in assorted teaching-and-learning initatives in high school, and was a TA for MIT 6.071 and 6.101 as a graduate student.

Piotr Mitros: Art, Imaging, and Photography
I do a bit of photography, videography, and occasionally, imaging equipment design. This section is coming!
Piotr Mitros: Academic Geneaology

This is a chart of my academic genealogy (showing thesis supervisor/supervisee relations). The data is primarily from the American Mathematical Society's Mathematics Genealogy Project. The chart was made with Graphviz, although I had to rescale the image map with a perl script. Note that it is an image map -- click on the people, and you'll go to their Wikipedia pages. The mystery is Seymour Papert. As far as I can tell, there is no record of who his advisor(s) in South Africa was/were.

Big chart showing my thesis advisors, their advisors, and so on. This chart is fundamentally graphical, but the link to the Math Genealogy Project above gives a screen reader friendly version..

Piotr Mitros: Education

Degrees

  • 2007 Ph.D Electrical Engineering, MIT. Advisors: Gerald Sussman and Tom Knight
  • 2004 M.Eng Electrical Engineering and Computer Science, MIT
  • 2004 B.S. Electrical Science and Engineering, MIT
  • 2004 B.S. Mathematics, MIT

As a bit geneaology, Jerry Sussman was co-advised by Seymour Papert. Based on my work on Open edX, I tend to get lumped with instructivist educational technology, but Papert's ideas on education inspire my work at least as much, as do many technology-free social constructivist concepts.

Other Certifications

  • MOR Associates Advanced Leadership Program
  • Scrum.org Professional Scrum Product Owner I
  • SSI Open Water Diver
  • NauticEd: Qualified Crew Member (Navigation Rules Clinic, Skipper Course, Anchoring a Sailboat Clinic, Basic Sail Trim Course)
  • MIT Sailing: Swim, Provisional, Firefly.
  • Community Boating: Kayak, Mercury, Windsurfing. NauticEd:

MOOCs on Teaching and Learning

  • Applying Principles of Behavior in the K-12 Classroom (Coursera)
  • Coaching Teachers: Promoting Changes that Stick (Coursera)
  • Early Childhood Education (Open2Study)
  • Foundations of Teaching for Learning 3: Learners and Learning (Coursera)
  • Foundations of Teaching for Learning 6: Introduction to Student Assessment (Coursera)
  • HarvardX GOV1368.1x: Saving Schools, Mini-Course I: History and Politics of U.S. Education (edX)
  • HarvardX GOV1368.3x: Saving Schools Mini-Course 3: Accountability and National Standards (edX)
  • HarvardX GOV1368.4x: Saving Schools Mini-Course 4: School Choice (edX)
  • HarvardX GSE2x: Leaders of Learning (edX)
  • ICT in Primary Education: Transforming children's learning across the curriculum (Coursera)
  • Instructional Methods in Health Professions Education (Coursera)
  • Surviving Your Rookie Year of Teaching: 3 Key Ideas & High Leverage Techniques (Coursera)
  • Teaching Character and Creating Positive Classrooms (Coursera)

MOOCs on Business and Law

  • HarvardX HLS2x: From Trust to Promise to Contract (edX)
  • Human Resources (Open2Study)
  • MITx Entrepreneurship 101: Who is your customer? (edX)
  • Management for a Competitive Edge (Open2Study)
  • Managing People: Engaging Your Workforce (FutureLearn)
  • Negotiation and Conflict Resolution (Open2Study)
  • Principles of Project Management (Open2Study)
  • Strategic Management (Open2Study)
  • Subsistence Marketplaces (Coursera)

MOOCs on Engineering

  • Pattern Discovery in Data Mining (Coursera)
  • Sports and Building Aerodynamics (Coursera)
  • TUMx AUTONAVx: Autonomous Navigation for Flying Robots (edX)
  • Tenaris Steel101x: Introduction to Steel (edX)
  • UPValenciaX: Dynamics and Control (edX)

Other MOOCs

  • HarvardX SW12.1X: China (Part 1) - The Political and Intellectual Foundations of China (edX)
  • Ideas from the History of Graphic Design (Coursera)
  • Questionnaire Design for Social Surveys (Coursera)
  • The Emergence of the Modern Middle East - Part I (Coursera)
  • The Horse Course: Introduction to Basic Care and Management (Coursera)

More Interesting Classes

EE: Advanced Circuit Techniques (6.331), Subthreshold Analog VLSI (6.973), Analysis and Design of Digital Integrated Circuits (6.374), Design of Analog MOS LSI (6.775), Integrated Microelectronic Devices (6.720), Solid State Circuits (6.301), Computer Architectures (6.823), Unusual Computer Architectures (6.911). (Undergrad coursework not listed)

Graduate signals, systems, and controls coursework: Feedback Systems (6.302), Wavelets and Filterbanks (18.327), Dynamic Systems and Control (6.241) (Undergrad coursework not listed)

Graduate CS coursework: Seminar in Theoretical CS (18.419), Approximation Algorithms (6.891) (Undergrad coursework not listed)

Mathematics: Discrete Math (18.062), Analysis I (18.100), Analysis II (18.101), Fourier Analysis (18.103), Linear Algebra (18.700), Algebra I (18.701), Algebra II (18.702), Topology (18.901), Grad Probability (6.431) (Common calculus classes not listed)

Physics: Waves (8.03), Statistical Mechanics (8.044), Introductory Special Relativity (8.20), Advanced Classical Mechanics (8.351J), Quantum Computation (18.435J) (Intro physics not listed)

In addition, I have taken a number of classes in economics, a couple in business and technology law, and a couple on crosscultural communications.

Expired Certifications

  • Amateur radio operator
  • CPR
  • Piotr Mitros: Personal

    I was asked why I don't have more personal information on my web page. I do have quite a bit, actually. You just need to find it. For the most part, my life consists of writing code, starting businesses, making things, traveling, trying to make the world a better place, and spending time with my friends and family. I don't say too much about my friends and family (I respect their privacy), but otherwise, all of that is pretty well represented here. And in many cases, where I went, what I built, and why does tell a lot of my story and who I am.

    And to keep in mind I have a dry sense of humor. To the best of my knowledge, all parts of this page are true. However, a few are not quite serious (and indeed, one or two as social commentary). A few of the not-so-serious ones only come out on a close reading or a very close reading. Which, I suspect, no one will ever do. But then, it is my personal web page, not yours.

    About this page

    Ah, you're looking at the HTML of this page? Fine choice. A person of style.

    This page is intentionally a very simple web page. It uses just CSS and HTML5. Much of it is auto-generated (mostly from my CV, my github repos, as well as projects from older pages). It has just three lines of JavaScript, so if you go to the page with a non-existent anchor, it will take you to the bio page. My previous page was HTML2. I find HTML2/HTML5 more elegant to work with than 3/4 or JavaScript.

    (I do permit a little bit of JavaScript for the country map; not mine own)

    And a lot of the page, also intentionally, follows SGML rather than XML, a la HTML2. In particular, SGML specifies rules — which browsers seem to follow — about omitted closing tags for elements such as paragraphs and list items. That just makes for cleaner, more human-friendly text.

    Right now, it needs styling. It's very, very ugly, unfortunately. I went with defaults of:

    • The serif font is Merriweather. The san-serif is OpenSans. Both (approximately) match. They have a similar x-height (Merriweather is, unfortunately, a little bit higher). Merriweather is mch more modulated, and a little bit heavier. I'd like a slightly lower x-height. Even though it's for display use, the high x-height takes away from the refinement of the page. If I have a chance, I'll find better matching fonts. The look I want would be best suited to a transitional serif such as Baskerville or Caslon, to give a more academic look, with a matching sans serif.
    • The colors are from a generic color matching tool. I'd like classic colors -- very sapia, or old document looking. That's not quite what I have, but it's the best the tool could do. If I went into old parchment colors, the tool would spit out pretty wild ones for complements.
    • The autogeneration also leads to some style inconsistency. It's not fundamental — it could be fixed with more/better logic in the autogeneration.
    • And there is still a11y, responsiveness, etc. fixes possible

    I should also sort out a lot of the obsolete/undergrad stuff from the current more interesting things.