Things I build for myself or for the open commons — distinct from client engagements. Each of these exists whether or not anyone pays for it, which is the whole point: they're the bench I think on, the substrate I sharpen judgment against, and the place where I get to argue with my own design choices in public.
Active
trading-box — algorithmic trading research bench
A concrete answer to the question "does any of this AI/ML actually help when there are real consequences for being wrong." Python / FastAPI / Postgres, walk-forward backtests across 53 crypto pairs and seven timeframes, hybrid baseline-plus-adaptive strategies with per-asset trend filters. The latest 3-year hybrid run reached ~95% of its own oracle ceiling — a stronger signal than I expected from a self-tuning system.
Not open source. The research notes are: I write up findings the way I'd want a collaborator to write them up for me — what was tried, what the numbers said, which assumption broke. Some of that surfaces on the blog over time.
fedorthinks.com — this site, in the open
Next.js 16 + FastAPI monorepo. Four locales (en/es/ru/zh), comments with owner moderation, MDX blog with RSS, JSON-LD structured data, per-slug OG cards, vanity subdomains that 301 to canonical paths. Built directing Claude Code on Opus — design and review are mine, almost no line of code typed by hand.
Source on GitHub. The interesting thing about reading the diff history isn't the code itself — it's seeing what an AI-native delivery loop looks like when the engineer's job is to specify, review, and decide rather than to type.
Method
The throughline across both projects: the engineer's job is shifting from typing code to specifying behavior, reviewing agent output, and owning the architecture. I think the best way to learn what that looks like in practice is to do it, in public, repeatedly, on real problems.
Coming
A few more things are in early form — an evaluation framework, a couple of small tools built sideways out of the trading work. I'll list them here when they're real enough to describe in one sentence without hand-waving.