I am currently looking for work!

Immensely curious computational chemist turned software developer. I am currently finishing my Masters' in dynamic languages interoperation. I mostly write Python, JavaScript, Scheme and C (roughly in that order).

My academic and work experience spans the stack from DevOps and sysadmin work (cloud and on-prem) to full-stack web apps to networking, medical imaging, web scraping, smart contracts, API reverse-engineering, compilers and interpreters, FFIs, computational chemistry, LLMs, microcontroller firmware, general/executive consulting, technical expert in litigation etc.

Most of my work is private and not available on this repo.

Happy member of the Mighty Gerbils 🐹

Je maîtrise le français et l'anglais. N'hésitez pas à entrer en contact dans l'une ou l'autre de ces langues!

marc-andre.belanger [-at-] umontreal.ca.

• B. Sc. in Chemistry, Université de Montréal
• M. Sc. in Computer Science, Université de Montréal (ongoing)

• Marc-André Bélanger and Marc Feeley, A Foreign Function Interface between Gambit Scheme and CPython. In Scheme Workshop (SW'22), September 2022. (pre-print)
• Marc-André Bélanger and Marc Feeley, A Scheme Foreign Function Interface to JavaScript Based on an Infix Extension. In European Lisp Symposium (ELS'21), May 2021.
• Ernzerhof, Matthias, M. Bélanger, Didier Mayou and Tahereh Nemati Aram. Simple model of a coherent molecular photocell. The Journal of chemical physics 144 13 (2016): 134102 .

• Marc-André Bélanger and Marc Feeley, Interfacing with CPython from Gambit Scheme. In 1st Languages, Architectures, and Tools for Heterogeneous Computing (LATHC) Workshop 2023, Februrary 2023
• Marc-André Bélanger and Marc Feeley, A lightweight approach for accessing Python modules from Gambit Scheme (Lightning Talk). In Scheme and Functional Programming Workshop (SFPW'21), August 2021.

• 2022–2025 | Software consultant

  Worked on various contracts for clients involving web application development, web scraping, consulting, API reverse-engineering, data collection, analysis, and sanitization for SMBs.

• 2024 | LLM Optimization Lead – Early Stage Startup

  Worked as an R&D contractor for a generative AI PaaS on LLM-based workflow optimization and evaluation. Full-stack involvement from cutting-edge prompt engineering to computation graph optimizations to JITing prompt optimizers using meta-prompting agentic techniques with DSPy and TextGrad.

• 2023–2024 | Lecturer – Université de Montréal

  I taught two introductory programming courses to hundreds of students. The course uses Python and the codeBoot environment, which I help build and maintain. I managed around 15 TAs for grading, assignment and lab hours. My teaching was highly rated by students.

• 2014–2023 | Software Developer and Consultant – Imagerie des Pionniers

  Built business-critical internal and customer-facing full-stack web apps for cloud and on-premise deployments. Acted as a consultant for stakeholders and decision-makers for procurement, strategy and litigation.

  I have built and maintained:

  • A web-based appointment scheduler for patients handling over 20 000 appointments a year
  • An SMS-based waiting room management system to contact patients just-in-time for their exam handling hundreds of patients daily for over 100 000 business interactions per year
  • An issue tracking and ticketing system for internal use
  • A DICOM image sharing portal for practicioners and patients

• 2016 | Technical Support Specialist / Software Developer – La Presse

  Built Apple device monitoring and reporting system using Apple Profile Manager. Created dashboards for management and tech support using PowerBI. Provided general IT support in a high-volume media environment.

• 2013–2015 | Teaching Assistant – Université de Montréal

  I was a teaching assitant for multiple undergraduate chemistry courses (general chemistry, physical chemistry, inorganic chemistry). My work was highly rated by students and innovative for the time, providing extensive course notes, interactive presentations using electronics structure calculations and other software for the benefit of the students.

In no particular order or categorization.

• codeBoot - Browser-based IDE for teaching programming

  One of the main contributors to the project. I have worked on every part of the project, from the front-end to the compiler to the backend services (cloud and on-prem). I have used codeBoot to teach introductory programming courses at Université de Montréal.

  See pyinterp below for details on the codeBoot Python <-> JavaScript FFI.

• jsrna – JavaScript RNA Base Pair Annotation

  Web-based RNA base-pair annotator/viewer written in JavaScript. Also created a novel annotation method using a ResNet-based transfer learning model. I reformulated base-pair annotation as a classification task (similar to MNIST) using a canonicalized projection method to create an alphabet. Novel, highly efficient and accurate method yet unpublished.

• BLINX - ESP32-based board for teaching programming

  I am a co-founder of the company that markets and sells the BLINX hardware and software. BLINX is a small board used for teaching programming. It can be plugged into a USB port and programmed from the codeBoot environment. It accepts a multitude of Grove sensors. I wrote a high-performance firmware proof-of-concept using ESP-IDF as part of our R&D.

• py2js – Python to JavaScript Compiler

  Wrote a compiler targeting a subset of Python 3. Used in production as part of the codeBoot project, a web-based Python IDE for teaching.

• pyinterp – CPS-Style Python Interpreter in Python

  Implemented the FFI layer as well as Python builtins, critical to codeBoot’s backend. The FFI integrates ideas from my research into a production system. The FFI allows users to seamlessly import and use JavaScript libraries from Python code. This requires careful grammar, tokenizer and parser extensions along with a well designed runtime to preserve the illusion of writing pure Python code.

  Try the FFI at https://codeboot.org by executing \alert("Hello!") at the REPL.

• codeBoot on Microcontrollers

  Implemented WebSerial communication between codeBoot and MicroPython. Migrated legacy development environments to Docker for higher reliability and reproducibility. Handled firmware build automation. Deep Wireshark sessions to debug a microcontroller WebSocket implementation.

• Ribbit Scheme

  Implemented the Ribbit VM in Haskell. Built classical source-to-source transformations (e.g., closure conversion) into the compiler. Bootstrapped an x86-64 backend. Worked on porting Racket’s “sets-of-scopes” expander to Ribbit.

• Gambit Scheme

  Extended Gambit’s infix reader for JavaScript and Python syntax. Built an interface between Gambit Scheme and CPython using the CPython C API. Built an interface between Gambit Scheme and JavaScript which you can try here (write \alert("Hello!") at the REPL).

• 2023 DIRO Hackathon

  Helped organize the Université de Montréal’s 2023 DIRO Hackathon. Participants built games using the codeBoot platform to run isomorphically on the browser and an ESP32-based board.

• CLA – Article Reading Club

  Organized a scientific article reading club for compiler and programming language theory groups. Also hosted a CTF for the group.

• Ethereum transaction graph explorer

  Back when these things were rare, I built an Ethereum transaction graph explorer using Neo4j. I used this to find evidence of wrongdoing in some personal investments.

• Making EVM smart contracts lie

  Wrote a PoC EVM smart contract on the testnet to exploit out-of-band problems on a major chain explorer which resulted in spoofed token metadata, misguiding potential investors.