William Berman
Machine Learning Engineer In Residence at Sutter Hill Ventures
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolWilliam Berman is a machine learning engineer in residence at Sutter Hill Ventures with 11 years of software engineering experience across startups and developer-focused AI firms. He shipped diffusion model work at Hugging Face—contributing VQ-diffusion integration and conversion tooling to the widely used diffusers library—and has a track record building production systems at Paradigm, Coinbase, and other engineering teams. Comfortable bridging research and product, he combines hands-on model engineering (attention, positional embeddings, VQVAE integration) with practical software delivery for crypto, risk-assessment, and infrastructure projects. A UC Santa Barbara CS graduate (3.9 GPA) based in the Bay Area, he pairs strong academic foundations with open-source impact and an aptitude for translating research prototypes into reusable tooling.
11 years of coding experience
7 years of employment as a software developer
Bachelor’s of Science, Computer Science, 3.9 GPA, Bachelor’s of Science, Computer Science, 3.9 GPA at UC Santa Barbara
English
Github Skills (7)
transformers10
diffusion-models10
pytorch10
machine-learning10
deep-learning10
python9
image-generation9
Programming languages (12)
Github contributions (5)
huggingface/diffusers
Nov 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:764 reviews, 12 commits, 255 PRs in 2 months
Contributions summary:William contributed significantly to the VQ-diffusion project, focusing on integrating and adapting VQ-diffusion models within the Diffusers library. Their work included adding support for VQ-diffusion's VQVAE and transformer components, implementing attention mechanisms and positional embeddings for discrete inputs. Additionally, the user developed a conversion script to port models from VQ-diffusion to the diffusers framework, streamlining model integration and utilization.
pytorchartdeep-learningimage2imagestate-of-the-art
williamberman/FlexFlow
Sep 2022 - Apr 2023
A distributed deep learning framework that supports flexible parallelization strategies.
Contributions:127 pushes, 5 branches in 6 months
pytorchflexibledeep-learningmachine-learningdistributed-computing
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.