.png){kind=logo}
CC
Chenghui Cai
Director of Applied Sciences specializing in reinforcement learning and agentic AI for finance
New York City, NYDirector of Applied Sciences16 years experienceDirectorArtificial IntelligenceFinancial ServicesFinTech
ScreenedIdentity Verified
Connect with Chenghui
Chenghui already has a relationship with Reval, so a warm intro from us gets a much better response than cold outreach.
Recommended
Already have an account?
About
Embodied AI/robotics ML engineer with hands-on experience deploying POMDP-based reinforcement learning controllers on real mobile robots and vehicle fleets. Strong in sim-to-real robustness (domain randomization) and production rollout practices (HIL, shadow-mode, canaries, safety instrumentation), and has published related work (mentions a NeurIPS paper).
Experience
Director of Applied SciencesAyata – Prescriptive Analytics
Co-FounderShenzhen Qianhai Jinxin Big Data Finance Service LLC
Staff Scientist / Analytics ManagerOpera Solutions
Co-FounderMQT Investments LLC
Quantitative ResearcherCantor Fitzgerald
Postdoctoral Research AssociateDuke University
Education
Duke Universitydoctorate, Mechanical Engineering (Intelligent Systems) (2008)
Tsinghua Universitymaster, Mechanical Engineering (2003)
Tsinghua Universitybachelor, Mechanical Engineering (2000)
Key Strengths
- Deployed POMDP/RL policies on physical mobile robots (sim-to-real)
- Robust sim-to-real transfer via domain randomization (noise, friction, delays, partial observability)
- Built rigorous verification pipeline: simulation stress tests → HIL → shadow-mode → staged autonomous rollout with human override
- Productionization of ML for robotics: distillation + quantization + hardware bench validation
- High-impact architecture decision-making: pivoted from RL-only to hybrid model-based + learned control, cutting latency ~50% and improving reliability
- Deep understanding of embodied AI challenges (modality gap, temporal coherence, distribution shift, safety, real-time constraints)
- Applied generative models to create synthetic data to improve training/information extraction
Discover more candidates like Chenghui
Search across thousands of pre-screened, high-quality, high-intent candidates on Reval.
Search Talent