Trading Platform
A private financial-systems project exploring real-time market data, software infrastructure, and the engineering behind modern markets.
- Type
- Private Project · Financial Systems · In Development
- Status
- Phase 4 · Low-Latency Engine
01Why I started it
It started with trading. It became about what's underneath.
I got interested in financial markets the way a lot of people do: by trading. After some wins and some losses, the part I couldn't stop thinking about wasn't the trades. It was what was actually happening underneath the interface.
That pulled me into market structure, financial infrastructure, real-time data, and the engineering behind modern trading systems. This project is where I'm working through all of that by building, one phase at a time.
02What I've built
Four phases. Three built, one in progress.
- 01
Backtesting engine
A reproducible Python engine for testing ideas against historical data, with the research kept separate from the code that runs it.
- 02
Forward paper test
The frozen system now runs forward on data that didn't exist when it was built. Paper only: no orders, no credentials, no money.
- 03
Dashboard
A read-only web app that rebuilds from the logs on every push, so it can show a number but never change one.
- 04
Low-latency engine
A Rust market-data engine that consumes a public exchange feed, replays deterministically, and measures its own latency at every stage.
03Current status
04
Low-Latency Engine
The backtesting engine, the forward paper test, and the dashboard are built and running. I'm finishing the fourth phase: a low-latency market-data engine in Rust, now in its second milestone. Every change is measured before and after, and the engine still places no orders and holds no credentials.
- Backtesting engine
- Forward paper test
- Dashboard
- Low-latency enginenow
04What I'm learning
Most of what it's teaching me applies well beyond markets.
How to think about real-time systems, where the input never stops and every shortcut eventually shows up as a bug. How to measure performance instead of guessing at it, and how to be suspicious of a good number. What financial data actually looks like up close, and how much work it takes to make it reliable.
It's also where I learned Rust properly, after doing the research side in Python, and where I got in the habit of treating a project as research: form a question, build the smallest thing that can answer it, look at the result honestly, and go again.
This is an active private project. Some technical, research, and strategy details are intentionally not shown publicly.
