Patrick Von Platen
Research Engineer at Mistral AI
France
Summary
🤩
Rockstar🎓
Top SchoolPatrick Von Platen is a Research Engineer at Mistral AI with eight years of experience building open-source ML infrastructure across transformers, diffusion models, speech, and multimodal LLMs. Previously Technical Team Lead at Hugging Face, he created and maintained the Diffusers library (18k+ stars, ~50k daily pip installs) and led the Audio team that helped popularize Wav2Vec2 and XLS-R. He combines low-level inference and systems work with model engineering—contributions include core Mistral inference changes, integrating Mistral and Pixtral into vLLM (tokenizers, vision encoder and patch merging), and performance improvements in accelerate and transformers. He also added safety-checking and inpainting features to Stable Diffusion and strengthened evaluation and dataset tooling, reflecting a knack for shipping practical, production-ready ML features. Based in France and grounded in doctoral research at Cambridge, he is known for making large models faster, memory-efficient, and safer in real-world deployments.
9 years of coding experience
3 years of employment as a software developer
Engineer’s Degree, Engineering, GPA: 3,71, Engineer’s Degree, Engineering, GPA: 3,71 at Ecole Centrale ParisDoctor of Philosophy - PhD, Computer Engineering, Doctor of Philosophy - PhD, Computer Engineering at University of CambridgeBachelor’s Degree, Business Administration and Engineering: Electrical Power Engineering (B.Sc.), 1.7 (German grading system), top 5%, Bachelor’s Degree, Business Administration and Engineering: Electrical Power Engineering (B.Sc.), 1.7 (German grading system), top 5% at RWTH Aachen University
English, French, German, Spanish
Stackoverflow
82reputation
2kreached
4answers
2questions
Github Skills (62)
transformers10
git10
pytorch10
contentmanagement10
distributed-training10
python10
evaluation10
image-processing10
testing10
diffusion-models10
machine-learning10
audio-processing10
model-management10
stable-diffusion10
inference10
Programming languages (10)
Github contributions (5)
mistralai/mistral-inference
May 2024 - Mar 2025
Official inference library for Mistral models
Role in this project:
ML Engineer
ML EngineerContributions:4 releases, 21 reviews, 35 PRs in 10 months
Contributions summary:Patrick primarily contributed to the `mistral-inference` repository, which focuses on inference for Mistral models. Their contributions involved significant code changes within the core inference library, including modifications to model architecture, attention mechanisms, and the integration of LoRA (Low-Rank Adaptation) for model fine-tuning. The user also added code related to vision-language adapters and pipeline parallelism, indicating involvement in optimizing inference and expanding model capabilities. These changes suggest a focus on improving the performance, features, and adaptability of the Mistral models for various inference tasks.
llmllm-inferencemistralai
huggingface/diffusers
May 2022 - Jan 2023
🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX.
Role in this project:Backend Developer & Machine Learning Engineer
Backend Developer & Machine Learning EngineerContributions:36 releases, 5419 reviews, 752 commits in 8 months
Contributions summary:Patrick's contributions primarily focus on implementing and documenting changes related to the Latent Diffusion model. Their work included adding pipeline documentation, logging improvements, and enhancing the stability of the inference steps. They also focused on applying attention slicing and adding inpainting capabilities to improve the model's performance.
pytorchartdeep-learningimage2imagestate-of-the-art
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.