Hongkai Dai
Founding Member Of Technical Staff at Stealth Startup
Palo Alto, California, United States
Summary
🤩
Rockstar🎓
Top SchoolHongkai Dai is a founding member of technical staff and robotics-focused engineer with 12 years of experience building model-based and physical AI systems from research to product. He holds an MIT PhD in EECS and spent nearly a decade at Toyota Research Institute advancing robotics perception, planning, and verification before joining a stealth Physical AI startup in Palo Alto. A hands-on backend developer and researcher, he has contributed to high-profile open-source projects like Drake and the Flexible Collision Library, improving SDP formulation, collision/distance algorithms, and solver interoperability. His work blends algorithmic rigor with engineering pragmatism—adding dual-solution support, memory-safe implementations, and thorough unit tests that fixed subtle GJK/EPA issues. Comfortable moving between deep math and production code, he’s known for surfacing corner-case correctness fixes that materially improve downstream system reliability.
12 years of coding experience
9 years of employment as a software developer
PhD EECS, PhD EECS at Massachusetts Institute of TechnologyExchange Electronic and Computer Engineering, Exchange Electronic and Computer Engineering at University of Illinois Urbana-ChampaignHong Kong University of Science and Technology (HKUST)
Stackoverflow
2,641reputation
22kreached
112answers
2questions
Github Skills (17)
algorithm10
algorithms10
c-language10
solver10
collision-detection10
mathematical10
data-structure10
eigen10
data-structures10
cprogramming-language10
modeling10
unit-testing9
optimization6
mathematical-optimization6
drake6
Programming languages (9)
Github contributions (5)
RobotLocomotion/drake
Feb 2014 - Jan 2023
Model-based design and verification for robotics.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:2142 reviews, 1364 commits, 1494 PRs in 9 years
Contributions summary:Hongkai implemented a constructor and related code for the SdpaFreeFormat class for the Drake robotics framework, specifically for the formulation of semidefinite programs. These changes included incorporating the ability to parse second-order cones, support for the retrieval of dual solutions, and the generation of SDPA files for compatibility with various solvers. The user's work also included improvements in memory management related to the handling of variables.
model-basedmodel-based-designroboticsdrakeverification
flexible-collision-library/fcl
May 2018 - Jan 2020
Flexible Collision Library
Role in this project:Backend Developer
Backend DeveloperContributions:4 reviews, 11 commits, 21 PRs in 1 year 8 months
Contributions summary:Hongkai primarily focused on improving the Flexible Collision Library (FCL), specifically in the area of distance and collision detection. They addressed and resolved issues within the GJKDistance function, including variable overload and EPA implementation. These changes, including the addition of comprehensive unit tests, demonstrate a focus on improving code correctness and robustness within the collision-detection framework.
collision-detectionmachine-learningcollisionpathfindingflexible
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.