EE by degree, markets by choice. Derivatives, volatility, execution — I take risk deliberately, learn whatever a problem demands, and get to a working build fast. Theory only counts once it runs on live data.
Every project here starts from a trading or allocation question, not from a framework I wanted to try.
A lock-free, low-latency market-data pipeline running against live exchange data.
| Queueing | Lock-free SPSC ring buffer |
| Data path | mmap zero-copy |
| Timing | rdtscp cycle-accurate instrumentation |
| Measured latency | 7.4µs p99 queue transit across 42K live samples |
| Feed | Bybit BTC perpetual futures, V5 API |
| Threading | Pinned threads |
| Analysis | Python and Plotly |
Includes order book reconstruction. Taken from skeleton to a documented p99 on live data over a multi-week debugging cycle: V5 migration, the SPSC buffer, the mmap writer, then the measurement layer.
Dynamic programming applied to energy arbitrage under a regime-switching market.
| Method | Bellman backward induction |
| Market model | Regime-switching |
| Result | 96% improvement over random baseline |
| Alpha | +31.71 against -64.46 random |
| Documentation | Whitepaper-style README plus companion notes on dynamics, environment, simulator, and solver |
Exotic Derivatives | Options | Futures | Algorithmic Trading
A production university management platform: departments, subjects, classes, users, and live analytics.
| Stack | PostgreSQL on Neon, Express, React, Node |
| Interface | shadcn/ui with Refine |
| Auth | Better Auth, Google and GitHub |
| Infrastructure | Arcjet rate limiting, Cloudinary, Site24x7 RUM |
| Deployment | Vercel frontend, Railway backend |
| Systems | C++20, lock-free structures, zero-copy I/O, cycle-level timing, POSIX syscalls |
|---|---|
| Quant | Market microstructure, MDPs, regime-switching models, options |
| Backend | Node, Express, REST APIs, authentication, PostgreSQL |
| Frontend | React, TypeScript |
| Tooling | Docker, Git, Linux |
Currently sharpening SQL, low-level design, and data structures for the engineering track.
Leaving GitHub from here — Academic Hub, live · LinkedIn