Jeremy Hsu
  • Home
  • Work
    • Rime: ETF regime allocator
    • Fifth Sanctuary: gold volatility expansion
    • Kintoun: ORB microstructure research
    • Competitive robotics
  • About
  • Resume
    • Software engineering
    • Quantitative research

Jeremy Hsu

Top Rime Fifth Sanctuary Kintoun Robotics Experience Contact

Jeremy Hsu

Computer Science at Penn State. I build systematic trading strategies and the infrastructure that runs them.

Open to Summer 2027 SWE & quant internships

Scroll
Live · deployed April 2026

Rime

Rime is a daily allocator for leveraged equity ETFs. Each morning it splits capital between leveraged exposure, a volatility hedge, and cash, using a shallow regime tree with a macro overlay on top of it. It also has a circuit breaker that cuts leveraged exposure when volatility spikes. I wrote the research and the live system against the same decision function, which prevents model/live divergence.

Python · pandas · NumPy · Alpaca API · Streamlit
47
Pytest tests
14 yr
Simulated window
~44%
Avg gross exposure, simulated
Read the writeup →
Schematic · allocation states
Live · deployed January 2026

Fifth Sanctuary

An intraday volatility-expansion system on gold. It runs long and short as two separate sleeves, each sized by a macro state vector and filtered on its own. When I went looking at why the short sleeve was losing, it wasn't losing at random. It lost under a specific set of macro conditions, so I filtered it out of those conditions instead of tuning the parameters.

Python · C#/NinjaScript · MQL5 · FRED
22 yr
Simulated window
~18%
Time in market
42%
Win rate, simulated
Read the writeup →
Schematic · sleeve structure
Research · December 2025

Project Kintoun

Standard opening range breakout strategies get chopped up by false breakouts. This notebook tests whether waiting for a retest of the breakout level, plus a VWAP trend check and a volume gate, filters enough of them out to be worth it on NQ futures. The methodology, the math, the code and the full trade log are public, including the parts that don't work.

Python · event-driven simulation · public repo
1,062
Simulated trades
2022–25
Sample window
MIT
Licensed, source public
Read the notebook →
Simulated · 1,062 trades · 2022–2025
Archive · FTC #16468 · 2019 to 2025

Green Lemons

I started Green Lemons in 2019 and ran it for six seasons, growing it to fourteen members. I wrote the autonomous control software in Java: vision-based navigation, odometry, and path planning. OPR is a least-squares estimate of per-team scoring contribution, and the best objective measure of team strength.

Java · OpenCV · odometry · path planning · FIRST Tech Challenge
#1
OPR worldwide
8,000+
Teams
6
Seasons led
Read more →
Schematic · autonomous path
Background

Experience

Jan 2026 – present
Founder & Quantitative Researcher
Weaver Capital LP · West Chester, PA
Aug 2026 – present
Undergraduate Researcher
Locomotion in Biology and Robotics Lab, Penn State · under Dr. Baxi Chong
Aug 2025 – present
Quantitative Research Lead
Penn State Quantitative Finance Club
Jun 2024 – Jul 2024
Research Intern
Robot Perception & Navigation Group, University of Delaware
Jun 2019 – Apr 2025
Founder & Technical Lead
FTC #16468 Green Lemons Robotics
#1 OPR, worldwide
2024 FTC World Championships, out of 8,000+ teams
Dean’s List Finalist
FIRST Tech Challenge, individual student award
Top 100, United States
IMC Prosperity qualifying rounds
4th degree black belt
Taekwondo, certified by Kukkiwon in 2025

Coursework in applied regression, discrete mathematics, linear algebra, and data structures.

Full background →
Get in touch

Summer 2027

I'm looking for software engineering, quant dev, or quant research internships. I'm most useful where the modeling and the systems work are the same problem: research infrastructure, backtesting and simulation frameworks, execution systems, and data pipelines that have to be correct.

Engineering resume → Quant resume →

© 2026 Jeremy Hsu · jeremyhsu2025@gmail.com

Performance figures on this site are simulated backtest results unless explicitly labeled otherwise. Simulated results have inherent limitations and do not represent actual trading. Past performance is not indicative of future results. Nothing here is an offer to sell or a solicitation of an offer to buy any security, or investment advice.