Alex Trott
Senior Research Scientist at Databricks Mosaic Research
United States
Summary
🤩
Rockstar🎓
Top SchoolAlex Trott is a Senior Research Scientist with a decade of experience combining deep learning and neuroscience to make large models faster and cheaper to train. At MosaicML/Databricks he focuses on algorithmic efficiency for LLMs, contributing to projects like llm-foundry and Composer where he optimized BERT pretraining, warmup strategies, and memory/layout improvements. Previously at Salesforce he applied reinforcement learning to socio-economic simulations in the AI Economist framework, improving reproducibility and scenario flexibility. He is fluent in Python and Matlab, with a Ph.D. in Neurobiology from Harvard that informs a rigorous, experimentally grounded approach to ML research. Alex mixes hands-on engineering—benchmarks, bug fixes and feature additions—with principled research, bridging prototype code and production training pipelines. Colleagues rely on him to translate complex theoretical ideas into practical, well-tested training improvements.
10 years of coding experience
7 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Neurobiology, Doctor of Philosophy (Ph.D.), Neurobiology at Harvard UniversityBachelor of Science (B.S.), Neuroscience, Bachelor of Science (B.S.), Neuroscience at Brandeis University
Github Skills (21)
transformers10
pytorch10
multi-agent-reinforcement-learning10
python10
machine-learning10
simulation-framework10
economics10
llm10
huggingface10
deep-learning10
trainings10
bert10
neural-network10
nlp10
modeling10
Programming languages (2)
Github contributions (5)
salesforce/ai-economist
Aug 2020 - Nov 2021
Foundation is a flexible, modular, and composable framework to model socio-economic behaviors and dynamics with both agents and governments. This framework can be used in conjunction with reinforcement learning to learn optimal economic policies, as done by the AI Economist (https://www.einstein.ai/the-ai-economist).
Role in this project:
Back-end Developer
Back-end DeveloperContributions:8 reviews, 14 commits, 3 PRs in 1 year 3 months
Contributions summary:Alex primarily focused on enhancing the functionality and robustness of the AI Economist framework. Their contributions included adding seed control for reproducible experiments, fixing typos, and refactoring the code for clarity. The user also added options to control the seed and expanded and adjusted the scenarios to support different agent setups and reward metrics. Further improvements included the expansion of the layout and the addition of a split world scenario.
behaviorsfairness-mlsimulation-frameworkmahjongdeep-reinforcement-learning
mosaicml/llm-foundry
May 2023 - Aug 2024
LLM training code for Databricks foundation models
Role in this project:ML Engineer
ML EngineerContributions:91 reviews, 20 PRs, 60 pushes in 1 year 3 months
Contributions summary:Alex implemented and benchmarked BERT pre-training and fine-tuning examples using the GLUE benchmark. Their contributions included support for Hugging Face models and the creation of a Mosaic BERT variant. The user made code changes to support MosaicBERT in 0.12.1 and addressed issues related to forward pass parameters. Furthermore, they added a starter script for single-task classification fine-tuning, streamlining the process for users.
deep-learningllmneural-networksnlppytorch
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.