Recurse Center
Participant
Selective programmers' retreat for self-directed research and learning. During my batch, I focused on building Raven, an open-source AI ecosystem bringing machine learning tooling to the OCaml language.
About Me
I enjoy building products and developer tools. After leading the OCaml Platform initiative at Tarides, I'm now focused on building Raven, an AI ecosystem for OCaml.
With Raven, I'm exploring what a modern ML ecosystem could look like in OCaml. The goal is to make scientific computing in OCaml as productive as Python, while leveraging OCaml's strengths in type safety and performance. I'm also exploring paths to sustainable development around this ecosystem.
Tarides
Director of Engineering/Engineering Manager
Tarides is on a mission to make software development more secure through OCaml.
I started as Principal Engineer and grew into leading 6 engineering teams as we scaled from a small tooling team to 40+ engineers. My main focus was the OCaml Platform - making OCaml's developer experience rival mainstream languages.
The highlight was launching the new OCaml.org. We went from a static site to a modern platform with package docs, job board, and playground.
I also drove our push to unify OCaml tooling through Dune package management and presented a community roadmap at ICFP 2023. I'm very proud of all the progress we made, shipping huge developer experience improvements across 10+ developer tools used by 400K+ developers.
Docent Art
CTO (Interim)
I was the interim CTO at Docent Art, where I built the technical foundation for an AI-powered art discovery platform. I designed and implemented a scalable infrastructure that became our competitive advantage: a data pipeline processing 300+ art websites daily, a crawling system, and a labeling platform for training our recommendation algorithms. The infrastructure allowed us to quickly prototype both web and mobile versions of our platform, supporting our product-market fit exploration and investor and client demos.
Keatext
NLP Engineer
Joined to lead the development of Sapiens, a data-labeling platform I had built independently.
As part of the R&D team, I also implemented and trained state-of-the-art NLP models including BERT and ELMo.
See more
Roof AI
Data Scientist
I worked as a Data Scientist at Roof AI, where I designed and built a sophisticated ETL pipeline that aggregated data from 20+ real estate servers (all using different data formats, RESO wasn't a thing yet). I contributed to our migration to Google Cloud and created business intelligence reports to optimize product conversion rates.
PatSnap
AI Research Developer
Early 2017, I moved to China to work at PatSnap, where I was the first member of the computer vision R&D team. I built algorithms to detect and extract data from patent images that powered our image-based search engine. This is where I started learning about Deep Learning. I first built algorithms using traditional CV methods, which we then converted to Deep Learning models later on.
We built an object detection model using Faster R-CNN that achieved 0.95 mAP (it was really good at the time!) on patent image data extraction. I also developed a framework and desktop application for patent data labeling that was used to label all our training data.
Mirametrix
Software Developer
As a Software Developer at Mirametrix, I developed computer vision algorithms for gaze and face tracking in C++ and Python. I created a framework for IR camera testing that was used to test 300K cameras.
I once was responsible for building a demo of sleep detection for an important car manuafacturer, I got to learn about game development using Unity, and had to work non-stop for a week to get the demo ready. It was a lot of fun!
École de technologie supérieure - ÉTS Montréal
B.Eng in Software Development
Part of SONIA (Système d'Opération Nautique Intelligent et Autonome), leading the software team
Stedis
System Administrator
While studying, I worked part-time as a System Administrator at Stedis (ex Egedis), where I modernized the IT infrastructure across three offices. I implemented centralized network management through LDAP, WDS, and WSUS servers, and migrated a 60+ computer IT park.
I had an unusual level of autonomy, so lot's of fun and many mistakes (I lost a user's emails once!), but I learned a lot and couldn't think of a better way to start my career.
Grenoble-Alpes University
B.Sc. in Networks and Telecommunication
Blog
Thoughts on building developer tools, open-source communities, and the future of scientific computing in OCaml.
Projects
Contributions
Talks
Publications
-
State of the OCaml Platform 2023
BibTeX
@inproceedings{2023_ocaml_platform, author = {Thibaut Mattio and Anil Madhavapeddy and Thomas Gazagnaire and David Allsopp}, booktitle = {proceedings of the 2023 OCaml Users and Developers Workshop}, month = {sep}, title = {State of the OCaml Platform 2023}, url = {https://icfp23.sigplan.org/details/ocaml-2023-papers/15/State-of-the-OCaml-Platform-2023}, year = {2023} } -
Real World OCaml (2nd
Edition) - The OCaml Platform
BibTeX
@book{real_world_ocaml_2nd, author = {Yaron Minsky and Anil Madhavapeddy and JD Unicorn}, title = {Real World OCaml: Functional Programming for the Masses}, year = {2024}, edition = {2}, publisher = {O'Reilly Media}, url = {https://dev.realworldocaml.org/} } - S.O.N.I.A. Autonomous Underwater Vehicle Concept and Design of the 2016 Prototype
Community Involvement
- Previously maintained OCaml.org, OCaml VSCode and OCaml LSP
- I started and organized ReasonMTL and OCaml Manila meetups
-
I have been mentoring Outreachy interns. Past internships include:
- OCaml.org Monitoring Dashboard by Jiae Kam (Project Repository - Watch Presentation)
- VSCode OCaml Extension (.eml support) by Sayo Bamigbade (Project Repository - Watch Presentation)