Griffin Bassman
Founder at GAB Capital
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolGriffin Bassman is a data scientist and founder with eight years of experience building production-grade ML and back-end systems, blending deep algorithmic work with practical engineering. He has driven reinforcement learning and contextual bandit improvements in Vowpal Wabbit and contributed to Microsoft’s Autogen framework, surfacing expertise in online learning, LLM integration, and model persistence. Trained at UPenn (BSE CS) and pursuing an MSc at UCL, he moves fluidly between C++, Python, Java and kdb+ to optimize performance-critical systems. Griffin’s recent roles span Microsoft Research and CoreAI, and he now runs a proprietary trading firm, applying research-grade ML to market-facing problems. Notably, his open-source commits emphasize robust model save/load, weight management, and reductions-based refactoring—small changes that materially improve reliability and efficiency at scale.
8 years of coding experience
5 years of employment as a software developer
Bachelor of Science in Engineering (BSE) Computer Science, Bachelor of Science in Engineering (BSE) Computer Science at University of PennsylvaniaHigh School Diploma, High School Diploma at The Browning SchoolUniversity College London
Github Skills (27)
algorithm10
algorithms10
c-language10
ai-agent10
python10
agentic10
elearning10
machine-learning10
implement10
agent10
bandit10
cprogramming-language10
reinforcement-learning9
application-framework9
refactor9
Programming languages (7)
Github contributions (5)
VowpalWabbit/vowpal_wabbit
May 2021 - Jan 2023
Vowpal Wabbit is a machine learning system which pushes the frontier of machine learning with techniques such as online, hashing, allreduce, reductions, learning2search, active, and interactive learning.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:3 releases, 645 reviews, 269 commits in 1 year 8 months
Contributions summary:Griffin primarily contributed to the core machine learning algorithms within the Vowpal Wabbit library, including implementations and improvements to various Contextual Bandit (CB) and Reinforcement Learning (RL) methods, particularly focusing on SquareCB, First, RegCB, and other related exploration techniques. The user's commits addressed functional aspects of these algorithms, such as adding save/load mechanisms for model persistence and ensuring correct weight management, thus enhancing model accuracy and usability. Further contributions include refactoring code to improve efficiency by leveraging reduction learners, and address build failures in the system.
hashingtechniquescpppythonactive-learning
microsoft/autogen
Mar 2024 - Mar 2025
A programming framework for agentic AI 🤖 PyPi: autogen-agentchat Discord: https://aka.ms/autogen-discord Office Hour: https://aka.ms/autogen-officehour
Role in this project:Back-end Developer
Back-end DeveloperContributions:10 reviews, 32 PRs, 132 pushes in 1 year
Contributions summary:Griffin primarily contributed to the back-end aspects of the project. Their commits show they fixed formatting issues, added new features, and addressed errors. The user modified code related to model configurations and added new models for OpenAI, suggesting an involvement with integrating and managing LLMs within the project. Additional contributions include implementing .NET exceptions and XML documentation, indicating efforts to improve code quality and interoperability.
agenticagentic-agiagentsaiautogen
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.