Robert Nishihara
Co-Founder at Anyscale
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolRobert Nishihara is a co-founder and seasoned software engineer with 13 years of experience building distributed systems and AI infrastructure from the Bay Area. He co-founded and led Anyscale, translating research-grade distributed runtimes into production-ready platforms that scale ML workloads. His open-source impact includes core contributions to Ray, improving stability and non-blocking fetches for the project’s distributed runtime, and enhancements to Apache Arrow’s Python bindings and Plasma object store for high-performance data interchange. Trained as a PhD computer scientist at UC Berkeley with a BA in Mathematics from Harvard, he bridges rigorous research and pragmatic engineering. His background includes research internships at Facebook AI Research and Microsoft Research and early systems experience at Google and Jane Street, reflecting deep expertise across ML systems, low-level performance tuning, and production reliability. Colleagues describe him as someone who turns subtle runtime issues into robust, observable systems enabling large-scale ML.
13 years of coding experience
6 years of employment as a software developer
Doctor of Philosophy (PhD) Computer Science, Doctor of Philosophy (PhD) Computer Science at University of California, BerkeleyBA Mathematics, BA Mathematics at Harvard University
Github Skills (33)
algorithms10
serializable10
c-language10
serializer10
system-programming10
python10
networking10
reinforcement-learning10
numpy10
scala10
serialization10
apache-arrow10
deserialization10
plasmoid10
performance-optimization10
Programming languages (7)
Github contributions (5)
ray-project/tutorial
Apr 2017 - Oct 2019
Role in this project:
Back-end Developer
Back-end DeveloperContributions:45 commits, 67 PRs, 22 pushes in 2 years 5 months
Contributions summary:Robert implemented exercises, including adding example models and updating exercises, and modifying rollouts.py with functions like rollouts, add_advantage_values, and collect_samples using NumPy and Ray for a reinforcement learning project. They also modified exercises to use the actor API. These actions involved manipulating reinforcement learning environments and implementing features using Python. The work indicates a focus on developing components for reinforcement learning and possibly building the underlying systems for Reinforcement Learning algorithms.
amplab/SparkNet
Nov 2015 - Apr 2016
Distributed Neural Networks for Spark
Role in this project:Backend Developer
Backend DeveloperContributions:133 commits, 63 PRs, 759 pushes in 5 months
Contributions summary:Robert contributed to the development of the `sparknet` project, which focuses on distributed neural networks for Spark. Their primary contributions involved implementing a new `NDArray` library in Scala, essential for handling numerical computations within the neural network framework. Additionally, they made improvements to existing code, specifically addressing whitespace issues.
neural-networkspark
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.