-3b82f6?style=for-the-badge&logoColor=white){kind=icon}
Private (PPL) → Instrument → Commercial → Multi-Engine → CFI → CFII / MEI → 1,500 hrs → ATP
The same habits that made me good at engineering — discipline, systems thinking, running the checklist right every single time — are exactly what the flight deck rewards. Follow the live progress at yashhooda.ai (real-time flight tracker included).
Training to become an airline pilot — currently in ATP Flight School's Airline Career Pilot Program at Sugar Land Regional (SGR), working toward my Private Pilot Certificate and, ultimately, the flight deck at a major carrier.
I got here as a Data & AI Engineer building production systems — the kind with rate limiters, kill switches, and a 3 a.m. pager, not the kind that lives in a notebook. That work is the foundation funding the flying, and the two are more alike than they look: both are about building reliable systems under real constraints and never trusting the happy path.
- Ship full-stack AI — multi-model LLM gateways, hybrid RAG with reranking, agentic loops, and vector search — all live in production on real traffic.
📍 Richmond, TX · 🎓 BS Computer Science, UT Dallas · 📜 Databricks Certified Data Engineer Associate
🧠 yashhooda.ai — flagship platform
Full-stack AI application · Node.js · Vercel · Upstash
Not a portfolio site — an AI application that happens to have my resume in it. Live ADS-B flight tracker, my pilot-training journey, a hybrid RAG chatbot (dense + sparse retrieval, RRF, CRAG grading, cross-encoder reranking), a multi-model gateway (Claude, GPT, Grok, Gemini, Llama), and a 7-layer security gateway. Survived a sustained real-world attack campaign — defenses held.
|
Python · FastAPI · Leaflet · Live API Real-time tracker for the Infinite Flight simulator: live map of every aircraft (coloured by flight phase), origin→destination cards with live ETAs, arrival weather + 5-day forecast, satellite/day-night layers, ATC frequencies, and pilot logbooks. FastAPI backend proxies and caches the Live API so keys never touch the browser. |
Python · pandas · scikit-learn · NOAA · GitHub Actions 56-year (1970–2026) Bronze→Silver→Gold climate pipeline across 13 global cities, auto-refreshed daily via GitHub Actions. Trend regressions, heat-day counts, and city comparisons — weather being a core pilot skill. Houston warming: +0.77°F/decade. |
🗺️ HoodaRoutesNext.js · FastAPI · Strava · Garmin Connect IQ Worldwide running-route generator: road-snapped loops calibrated to a target distance, personalized to your Strava history, with one-tap push to Garmin and a companion Connect IQ watch app sideloaded onto a Forerunner 970. |
Python · MCP · FastMCP · Docker Model Context Protocol server that lets Claude read Garmin activities and push structured workouts and full training plans straight to a watch. Reverse-engineered Garmin's workout-service schema into a typed, LLM-friendly spec — garth SSO auth, a bearer-secured transport, offline tests, and CI. |
AI / LLM Engineering
RAG (hybrid + RRF + reranking) · CRAG · Agentic loops / ReAct · QLoRA fine-tuning · MLflow · Evals · Vector search (Upstash, FAISS, Chroma)
Data Engineering
Medallion architecture · Incremental ETL with control tables · REST API ingestion · Idempotent replay · Data quality & reconciliation
Aviation — training toward the airline flight deck at ATP, and building flight-tracking tools for fun. Running — marathoner chasing sub-3:00, which is why half my side projects are running infrastructure. Also: weather systems, NBA, independent science, and open-ended research rabbit holes that end at 2 a.m. with a new repo.
Building AI that ships. Training to fly. Open to what's next.
