Anton Lozhkov
Machine Learning Engineer at Hugging Face
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolAnton Lozhkov is a Machine Learning Engineer with 8 years of experience specializing in making cutting-edge NLP and multimodal transformer research practical and efficient for production. Based in Paris, he builds and optimizes transformer-based models for speech, audio, and vision, with hands-on expertise in few-/zero-shot learning, diffusion models, and ML infrastructure. At Hugging Face he has contributed to flagship open-source projects like Transformers, Datasets and Diffusers—implementing Wav2Vec2 pretraining, integrating speech benchmarks, and adding GLIDE UNet and scheduler components—bridging research and real-world tooling. Previously he delivered transformer-based pipelines for multilingual and legal document processing, syntax/NER systems for Russian and English, and cross-lingual search at scale. Comfortable across model development, dataset engineering, and deployment, he often focuses on production-ready efficiency and low-resource workflows that aren’t obvious from model papers alone.
8 years of coding experience
4 years of employment as a software developer
Master's degree (incomplete), System and Software Engineering, 9/10, Master's degree (incomplete), System and Software Engineering, 9/10 at Higher School of Economics
English, Russian
Github Skills (34)
notebook10
pytorch10
wafv210
python10
diffusion-models10
machine-learning10
diffusers10
implement10
datasets10
huggingface-transformers10
deep-learning10
speech10
natural-language-processing10
transformer10
model-conversion10
Programming languages (3)
Github contributions (5)
huggingface/diffusers
Jun 2022 - Jan 2023
🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX.
Role in this project:
ML Engineer
ML EngineerContributions:12 releases, 397 reviews, 286 commits in 7 months
Contributions summary:Anton implemented and added various files related to the "glide" model, including UNet architecture, cosine schedule for the scheduler, and related configuration files, indicating a focus on diffusion models. They also converted model weights from a different format and added the text encoder for the "glide" model, suggesting involvement in model integration and adaptation. Furthermore, the user worked on a full UNet model with attention and timestep embedding, highlighting expertise in model development within the diffusion model domain.
pytorchartdeep-learningimage2imagestate-of-the-art
huggingface/notebooks
Oct 2021 - Jan 2023
Notebooks using the Hugging Face libraries 🤗
Role in this project:ML Engineer
ML EngineerContributions:4 reviews, 11 commits, 23 PRs in 1 year 3 months
Contributions summary:Anton primarily contributed to the development and maintenance of machine learning notebooks within the Hugging Face `notebooks` repository, focusing on utilizing Hugging Face libraries. Their work involved adding and updating notebooks for various tasks, particularly in the domain of diffusion models. The user's contributions included integrating and bumping diffusers library versions, showcasing their expertise in the application of these models. They also addressed dependencies and configurations necessary for training and running the notebooks.
nlpdeep-learningmachine-learninghugging-facenotebooks
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.