Denis
Ambrosov
AI Engineer
How it works
- 01
Learn
Dig into the papers, courses, and models until the idea actually clicks — not just the summary.
- 02
Build
Turn the theory into something that runs. Code, experiments, small systems that either work or don't.
- 03
Document
Write down what worked, what broke, and why — while it's still fresh and the scars are honest.
- 04
Publish
Ship it in public: posts and videos, so the next person can skip the dead ends I hit.
Two tracks
AI / ML
Concept explainers, ML system design, and interview/career prep — Python, deep learning, RL, agents, and paper breakdowns. Built and documented in the open.
Crypto & Prediction Markets
On-chain analysis, market microstructure, DeFi, and the mechanics of Polymarket / Kalshi — treated as education, never as personal trading disclosure.
Who I am
I work in model risk at a European bank. The rest of my time goes into AI/ML, learned the way that actually sticks for me: build something, write down how it went, publish it, then do it again. This site is the publish step.
Most of what I make is applied AI/ML: concept explainers, ML system design, and the interview prep I'm grinding through myself. There's a second track too, on crypto and prediction markets — microstructure, on-chain analysis, and how venues like Polymarket and Kalshi actually work. I keep that one as education, not trading advice.
I'm based in Thailand and recently got engaged. I'd rather show work than talk about it, so the good stuff ends up in the writing and videos in Labs.
Credentials
- FRM
- CFA Level 2 Candidate
- CMT
- PRM
- SOA — Exam P
- SOA — SRM
Teaching & mentoring
ITMO University
Guest lecture on LLM inference.
Karpov Courses
Mentoring aspiring ML practitioners.
Grow together
Newsletter
One email when there's something worth reading — new write-ups and videos, nothing else.
Newsletter coming soon — the sign-up opens with the first posts.
Follow along
DMs are open — the fastest way to reach me is on any of these.