John Miller
Member Of Technical Staff at Anthropic
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolJohn Miller is a machine learning researcher and engineer with 11 years of experience, currently a Member of Technical Staff at Anthropic after co-founding a proprietary trading firm focused on quantitative strategies. He holds a PhD in EECS from UC Berkeley and MS/BS degrees from Stanford, and his work spans robust ML systems for financial markets, energy systems, and language/speech research. John has contributed to high-profile open benchmarks like Google’s BIG-bench, implementing reading-comprehension tasks and baselines that reflect attention to reproducibility and evaluation rigor. Equally comfortable in research and production, he blends academic depth with entrepreneurial execution and a taste for real-world, high-stakes ML applications.
11 years of coding experience
5 years of employment as a software developer
Doctor of Philosophy - PhD, Electrical Engineering and Computer Science, Doctor of Philosophy - PhD, Electrical Engineering and Computer Science at University of California, BerkeleyMcCallie SchoolMaster of Science - MS, Electrical and Electronics Engineering, Master of Science - MS, Electrical and Electronics Engineering at Stanford University
Github Skills (5)
machine-learning10
nlp10
python10
natural-language-processing10
json9
Programming languages (4)
Github contributions (5)
google/BIG-bench
Mar 2021 - Jul 2021
Beyond the Imitation Game collaborative benchmark for measuring and extrapolating the capabilities of language models
Role in this project:
ML Engineer
ML EngineerContributions:20 commits, 1 PR, 5 comments in 4 months
Contributions summary:John's contributions primarily involve the development and extension of a reading comprehension task within the BIG-bench framework. They implemented the SQuADShifts task, a reading comprehension benchmark, and integrated it into the existing codebase. They added support for features like `max_examples` and `random_seed`, showing an understanding of the task's flexibility. Additional work includes adding baselines and code formatting.
bertmachine-learningbenchmarkmeasuringbenchmarks
modestyachts/squadshifts-website
Jul 2020 - Oct 2022
Contributions:3 PRs, 6 pushes, 2 branches in 2 years 3 months
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.