# Built a custom Rust HFT engine (2.66 Sharpe). How do I properly stress-test profitability before scaling? — retained record

> Retained reddit record from reddit:algotrading.

Canonical: https://highsignal.app/data/records/f8b26c68f94c70c1

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:algotrading)  published 2026-11-02 19:59 ingested 2026-11-03 02:43

[original source ↗](https://www.reddit.com/r/algotrading/comments/1xm7qfz/built_a_custom_rust_hft_engine_266_sharpe_how_do/)

## retained text



```
I’ve spent a long time building a low-latency execution engine in Rust. On paper and in clean simulations (48M RPS, 332ns internal latency), the metrics look solid: 85% win rate and a 2.66 Sharpe ratio.

Now I’m moving toward low-capacity live testing on micro-accounts, but I want to be smart about it. The market microstructure usually destroys perfect backtests.

What are the best ways to verify this profitability in a real live environment without blowing up the account on toxic flow or hidden execution penalties? Also, for those running similar setups, what features or risk-management layers are absolutely essential to add at this stage?
```



document fetched 2026-11-03 00:45 · canonical https://reddit.com/r/algotrading/comments/1xm7qfz/built_a_custom_rust_hft_engine_266_sharpe_how_do
