#ai
What happened after Part 2 published: the maintainer read it, opened an issue, shipped two fixes in 48 hours, then released v1.10.0 with every workaround from this series built in natively. The complete, current pipeline setup — and what a first open source contribution arc actually looks like.
How quantitative funds try to tell a real, repeatable pattern from a correlation that's just noise, what actually makes one quant fund better than another, and whether high-frequency trading is, on balance, good or bad for markets.
The complete, runnable build of the pipeline from part 1: fli + trvl + an AI agent, full install and MCP config, a reusable prompt template, why I run flights and accommodation sequentially, and how I get accommodation prices that are actually bookable instead of phantom teasers.
How I built a flexible flight + accommodation search pipeline using fli and trvl — without API keys.