Emrick Sinitambirivoutin
Core Team Engineer at H Company
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolEmrick Sinitambirivoutin is a machine learning and embedded-systems engineer with five years of experience bridging hardware acceleration and production AI, currently serving as a Core Team Engineer in Paris. He has progressed from FPGA and RTL development for deep learning to leading ML engineering efforts at Sonos, delivering optimized inference pipelines and quantized operator implementations. Emrick’s rare combination of FPGA, CUDA, and embedded-software expertise enables him to design solutions that run efficiently both in data centers and at the edge. Notably, he contributed quantized softmax support to the popular tract inference library, demonstrating practical low-level ML optimization skills. With academic exchanges at EPFL and a background in physics-electronics, he pairs rigorous technical foundations with hands-on deployment experience.
5 years of coding experience
6 years of employment as a software developer
Embedded systems and Software devices, Embedded systems and Software devices at National School of Computer Science and Applied Mathematics of GrenoblePhysics Electronics and Telecommunications, Physics Electronics and Telecommunications at Grenoble INP - PhelmaBaccalauréat scientifique Mention européenne Anglais, Baccalauréat scientifique Mention européenne Anglais at Lycée Gerville RéachePreparatory classes Mathematics physics, Preparatory classes Mathematics physics at Prépas BaimbridgeAcademic exchange Master Data Science, Academic exchange Master Data Science at EPFL
French, créole et pidgin, basés sur le français, English, Spanish
Github Skills (9)
neural-network10
quantization10
artificial-intelligence10
fixed10
rust10
tensorflow10
onnx10
fixed-point10
testing9
Programming languages (2)
Github contributions (5)
sonos/tract
Apr 2022 - Dec 2022
Tiny, no-nonsense, self-contained, Tensorflow and ONNX inference
Role in this project:
ML Engineer
ML EngineerContributions:9 reviews, 49 commits, 39 PRs in 7 months
Contributions summary:Emrick primarily contributed to the implementation of a softmax operation, a core component in neural networks, within the `tract` library. Their work involved adding support for quantized data types, crucial for efficient inference. The commits include modifications to the core softmax functionality, the introduction of fixed-point arithmetic for quantized operations, and test cases to validate the correctness of the implementation.
rust-libraryno-nonsenseself-containedrustdeep-learning
emricksinisonos/rumqtt
Oct 2020 - Aug 2024
A fast, lock free pure rust mqtt client
Contributions:2 pushes in 3 years 10 months
lockrustmqttmqtt-clienthome-assistant
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.