Julien Balian
Senior Machine Learning Engineer at Sonos, Inc.
Greater Paris Metropolitan Region France
Summary
🤩
Rockstar🎓
Top SchoolJulien Balian is a Senior Machine Learning Engineer based in Greater Paris with 7 years of experience building production-ready speech and audio ML systems for embedded devices and cloud services. He has led speaker recognition efforts and improved acoustic and keyword-spotting models at Sonos, bringing research-grade approaches (metric learning, prototypical scoring, clustering) to Raspberry Pi–class hardware. Julien founded a speech-analysis SaaS, designing end-to-end pipelines from data collection and large-scale training to BI dashboards and microservice orchestration, showing both product and engineering ownership. His background spans speech recognition, diarization, signal processing and NLP, with hands-on implementation in PyTorch, TensorFlow, Rust and scalable tooling on AWS. An active contributor to the tiny inference library tract, he has fixed low-level operator bugs and added support for new ops and data types—highlighting his ability to work across model research and inference-runtime details. With an MSc in Software Engineering (Distinction) and a track record of turning academic methods into deployable systems, he excels at squeezing ML performance onto constrained devices.
7 years of coding experience
6 years of employment as a software developer
4th year - Master's degree, Computer Science, 4th year - Master's degree, Computer Science at SUPINFO - The International Institute of Information TechnologyMaster of Science (MSc), Computer Software Engineering, with Distinction, Master of Science (MSc), Computer Software Engineering, with Distinction at Oxford Brookes University
English, French
Github Skills (8)
rust10
onnx10
tensorflow10
neural-network9
artificial-intelligence8
data-structure7
data-structures7
algorithms7
Programming languages (3)
Github contributions (5)
sonos/tract
Apr 2022 - Jan 2023
Tiny, no-nonsense, self-contained, Tensorflow and ONNX inference
Role in this project:
Backend Developer
Backend DeveloperContributions:1 review, 23 commits, 28 PRs in 9 months
Contributions summary:Julien primarily focused on implementing and refining features within the `tract` library. Their contributions involved fixing bugs related to the `tile` and `unstack` operators, including serialization and handling of stride. They also added new operators like `unstack` and `stack`, and incorporated type conversions. Furthermore, the user made improvements to the core logic and added support for new data types.
rust-libraryno-nonsenseself-containedrustdeep-learning
JulienBalianSonos/tract
May 2023 - Jan 2025
Tiny, no-nonsense, self-contained, Tensorflow and ONNX inference
Contributions:90 pushes, 20 branches in 1 year 8 months
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.