Jeremy Hsu
Jeremy Hsu
Selected work
A daily allocator that splits capital between leveraged equity ETF exposure, a volatility hedge, and cash, using a shallow regime tree with a macro overlay and a volatility circuit breaker. Research and production call the same decision function, so live orders can't quietly drift from the backtest.
An intraday gold system running long and short as two separate risk sleeves, each sized by a macro state vector and filtered independently. The research found one sleeve was losing money structurally rather than randomly, which changed the design more than any amount of tuning would have.
A research notebook testing whether a break-and-retest microstructure filter plus VWAP trend confirmation can suppress false breakouts in Nasdaq-100 futures. Full methodology, math, and code are published, including where it doesn't work.
Founded and led a 14-member competitive robotics team across six seasons, writing the autonomous control stack, including vision-based navigation and path planning. Finished #1 OPR globally out of more than 8,000 teams at the 2024 FTC World Championships.
Experience
- Operate a multi-strategy systematic portfolio across futures and equities, spanning intraday, swing, and macro horizons.
- Engineered the execution infrastructure in Python and C#, with automated strategy synchronization, enforced risk limits, and sub-second order handling.
- Built portfolio optimization pipelines that allocate capital dynamically across uncorrelated sleeves and balance risk at the portfolio level.
- Researching locomotion and gait optimization in unconventional robot morphologies under Dr. Baxi Chong.
- Produce weekly market outlooks, technical analyses, and trade ideas for equity index futures.
- Developed forecasting models focused on macroeconomic cycles and market regime shifts.
- Built a 2D SLAM pipeline on a TurtleBot3 in C++ with ROS 2 and Google Cartographer, mapping an indoor lab environment.
- Integrated Nav2 for autonomous waypoint navigation on the resulting map.
- Presented navigation and sensing results to graduate researchers and faculty.
- Finished #1 OPR globally among 8,000+ teams at the 2024 FTC World Championships.
- Founded and led a 14-member team across six seasons; built the autonomous control software, including vision-based navigation and path planning.
Education
Pennsylvania State University, B.S. Computer Science, Aug 2025 – Dec 2028, University Park, PA
Technical
Awards
Get in touch
Best reached at jeremyhsu2025@gmail.com. Code on GitHub, background on LinkedIn, and the one-page version is here.