Guillaume Becquin
Team Lead - Search & Generative AI Applications
London, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolGuillaume Becquin is a hands-on AI and search engineering leader with eight years of experience building production-grade NLP, knowledge graph, and retrieval systems, now leading a Team of Research Engineers at Bloomberg to deploy LLM-driven search for the terminal. He blends applied ML research with strong software engineering — from prototype to scale — informed by prior roles as a deep learning solutions architect at Munich Re and data science work at Siemens and GE. Guillaume is an active open-source contributor to major ML projects, including performance and conversational pipeline improvements for Hugging Face Transformers and GPU/Windows support and serialization features for tch-rs (Rust bindings for PyTorch). Known for bridging research and product, he has shipped end-to-end SDKs and libraries, led cross-functional technical teams, and routinely navigates both model optimization and deployment complexities. His background in aerospace and engineering physics underpins a methodical approach to problem solving and system design.
8 years of coding experience
14 years of employment as a software developer
Dip. Ingénieur Aerospace Aeronautical and Astronautical Engineering, Dip. Ingénieur Aerospace Aeronautical and Astronautical Engineering at Ecole nationale supérieure de l'Aéronautique et de l'EspaceEngineer's degree Engineering Physics, Engineer's degree Engineering Physics at The Faculty of Engineering at Lund University
French, English, German, Spanish, Swedish
Github Skills (20)
pytorch10
language-model10
python10
machine-learning10
deep-learning10
model-optimization10
pipeline10
conversational-ai10
neural-network10
transformer10
nlp10
rust10
testing9
unit-testing9
gpu9
Programming languages (4)
Github contributions (5)
LaurentMazare/tch-rs
Dec 2019 - Dec 2022
Rust bindings for the C++ api of PyTorch.
Role in this project:
ML Engineer
ML EngineerContributions:2 reviews, 14 commits, 20 PRs in 3 years
Contributions summary:Guillaume significantly contributed to the `tch-rs` repository, which provides Rust bindings for PyTorch. Their work included implementing GPU support for Windows and adding a LayerNorm layer. Further contributions involved extending embedding configurations and adding unit tests for these new features, including support for f16 and i8 serialization. The user also refactored the var store and fixed the Windows LibTorch URL, demonstrating a strong understanding of the library's core functionality and build process.
pytorchapideep-learningrusttorch
huggingface/transformers
Dec 2019 - Jan 2023
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Role in this project:ML Engineer
ML EngineerContributions:4 reviews, 14 commits, 16 PRs in 3 years 1 month
Contributions summary:Guillaume contributed significantly to the Hugging Face Transformers library, focusing on improvements and additions related to conversational AI pipelines and model optimization. They implemented a new DialoguePipeline, refactoring it to a ConversationalPipeline, and integrated it with encoder-decoder models. Further contributions involved performance optimizations within the generation utilities, specifically regarding bad word handling and token masking, and also included bug fixes and test additions to support model functionalities.
pythonbertspeech-recognitionstate-of-the-artflax
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.