╔═══════════════════════════════════════════════════════════╗
║ AI SYSTEMS ENGINEER × PERFORMANCE ARCHITECT ║
║ Rust · WebAssembly · PyTorch · FastAPI · C++ ║
╚═══════════════════════════════════════════════════════════╝
Building at the intersection of AI infrastructure, systems programming, and full-stack product engineering. Not a web dev. Not an ML researcher. Something in between — and lower down.
- Rust → Wasm pipelines with
wasm-bindgen - Offloaded compute-heavy algorithms from JS → Rust/Wasm: 2–5× faster
- Reduced JS↔Wasm bridge calls by ~90% using shared memory buffers
- Async execution with Tokio runtime
- RAG pipelines — domain-grounded retrieval, hallucination reduction via verified sources
- Multimodal systems — vision + text + voice + image recognition, all wired together
- Vector infrastructure — FAISS, pgvector, semantic search, retrieval caching
- Forecasting models — Chronos-style time-series prediction
- Full AI system stack:
model → vector DB → RAG → backend API → client
- Production APIs with FastAPI + Node.js — not tutorial CRUD
- Microservices, async task queues, caching layers
- Docker, VPS, Linux server management, self-hosting
- Stripe payments, Mux video streaming, Firebase sync
- Unreal Engine 5 — C++ & Blueprints
- Unity — C#
- Wave spawning, gameplay mechanics, leaderboard systems
|
|
|
Email: pstanwar6747@gmail.com
Shipping compute-heavy systems — from the game loop to the inference pipeline.