18 · AI & Data Engineering @ Ankara University · building agentic trading infrastructure
I build production systems solo by orchestrating fleets of AI coding agents. My lane is the intersection of market microstructure and AI agents.
- polymarket-mcp-server — MCP server that gives Claude eyes on Polymarket: market search, order books, price history, and a liquidity-rewards scanner that ranks markets by reward pool ÷ on-book competition. 6 tools, zero credentials needed.
npx polymarket-mcp-server - hyperliquid-liquidation-toolkit — liquidation-price math and V-Risk cluster detection for Hyperliquid perpetuals. Pure Python, stdlib-only, 48 tests.
- ankaraunitenis.com (source) — production platform for Ankara University's 600+ member tennis community: league & tournament system, admin CMS, E2E-tested. Next.js 16 / Supabase. Built and operated solo, 279 commits.
Top 0.5% of ~2.6M candidates, Turkish national university entrance exam (2025). Built real-time trading systems against Hyperliquid's L2 order-book feed, a live Avellaneda-Stoikov market maker on Polymarket's CLOB, and backtesting engines with Monte Carlo + walk-forward validation — then killed the strategies my own data couldn't defend.
X: @birantx01 · LinkedIn: birant-egin
