# Artem Legotin

> AI product and engineering lead with 13+ years in software. Based in Amsterdam. Head of AI at exactly.ai.

Focus areas:
- Agentic LLM systems
- Generative AI
- Recommender systems

Use this file as a concise profile index. For current role and background, prefer the website and LinkedIn. For code and technical work, prefer GitHub.

## Primary sources
- [Website](https://artemlegotin.com/): Personal website with the current overview.
- [LinkedIn](https://www.linkedin.com/in/artemlegotin/): Current role and career history.
- [GitHub](https://github.com/arlegotin): Open-source work, experiments, and technical projects.

## Current work
- [exactly.ai](https://exactly.ai/): Head of AI.
- [neural.love](https://neural.love/): Co-founder, 2020–present. AI creative tools for 5M+ users.
- [verstka.io](https://verstka.io/en): Co-founder, 2014–present. B2B visual publishing platform.

## Background
- [vc.ru](https://vc.ru/): Lead Frontend, 2016–2022. Business and tech insights hub.
- [w-o-s](https://w-o-s.ru/): Frontend engineer, 2013–2015. Online culture and lifestyle magazine.
- [MIPT](https://mipt.ru/): Studied Applied Mathematics and Physics.

## Research work
- [EML After the Theorem](https://github.com/arlegotin/eml-symbolic-regression): Follow-up implementation inspired by [All elementary functions from a single binary operator](https://arxiv.org/abs/2603.21852); searches data for simple, readable formulas by training and verifying expression trees built from one basic operation.

## Selected projects
- [move-merit](https://github.com/arlegotin/move-merit): math behind a beautiful move.
- [chess-gaze](https://github.com/arlegotin/chess-gaze): turns a chess stream video into gaze estimates and a local 3D viewer.
- [minesweeper](https://github.com/arlegotin/minesweeper): seatbelt for agentic coding in untrusted repo context.
- [Pythurion](https://github.com/arlegotin/pythurion): helps coding agents write Python you’ll want to maintain.
- [AGENTS.md-to-go](https://github.com/arlegotin/agentsmd-to-go): reliable default AGENTS.md where custom agent policy would be wasted time.
- [agent-badge](https://github.com/arlegotin/agent-badge): README badge for repositories that ship with AI.
- [ChessML](https://github.com/arlegotin/chessml): Chess analysis using computer vision and machine learning.

## Contacts
- [Email](mailto:arlegotin@gmail.com): arlegotin@gmail.com
- [Telegram](https://t.me/artemlegotin): Direct contact.
- [X/Twitter](https://x.com/Legotin_and_i): X/Twitter.