Aly Thabet

I’m an AI engineer specializing in LLM orchestration and agentic systems, with a production-minded focus on reliability, cost, and guardrails. I build agentic workflows as deterministic, testable systems—so changes are validated before they touch real outcomes.

I designed and deployed APOLLO, an autonomous agentic trading system using the Anthropic Claude API as a multi-gate decision layer with live execution on Binance. It includes a 4-gate LLM pipeline (screen → analyze → confirm → size) plus a SmartExitManager with trailing take-profit, DCA safety orders, and strategy confluence scoring.

My approach centers on engineering discipline: Redis-backed state, a SQLite audit trail, idempotent order submission via CCXT, and careful rate-limit handling. I also treat prompt and policy updates like code by using backtesting-as-evaluation to run offline regression checks.

I’m currently seeking a junior AI language engineering role where shipped LLM-agent experience matters more than years on a CV. I bring hands-on experience from both applied ML work and backend-oriented infrastructure built with FastAPI, async patterns, and local inference tools.