Mathias
Tock.

context

Came in solo to build, from scratch, a complete data platform able to evolve with the coaching team's needs — not a one-shot dashboard but infra that earns its keep over seasons. The brief included custom analytical tools, the most distinctive being a fog-of-war simulator rebuilding what each side saw at any tick of a match.

approach

• Microservices architecture behind a single API gateway (auth, rate-limit, routing).
• Event-sourced Postgres writes where history matters; managed Supabase where it doesn’t.
• Observable by default — every hop traced, every job idempotent, no silent failures.
• Custom tooling: fog-of-war simulation, domain-specific aggregations, coach-facing views.
• Secure multi-tenant setup: JWT trust model, secret isolation, pre-merge security audit.

context

Cofounded a complete League of Legends coaching platform that turns each player's SoloQ history into a personalized learning track. Every match feeds 100+ metrics into a pipeline that picks the right video lesson and the right practice puzzle for the gap the data just revealed.

approach

• Cofounder — idea-first, then the build. Product and engineering owned by the same person.
• End-to-end scope on the frontend plus the data ingest + modeling layer.
• 100+ metrics extracted per SoloQ game, stored in Postgres, served with Redis-backed caches.
• Personalized coaching flows — video lessons + interactive puzzles, picked per-player, per-gap.
• Beta launched with real players who come back between games.