AI startups and high-concurrency backend teams building asynchronous Python APIs frequently search to hire a senior freelance FastAPI, Pydantic v2, SQLAlchemy 2.0 (`AsyncSession`) & `asyncpg` programmer. I am Aldian Fazrihady, a Senior Freelance Python & AI Application Engineer available via Upwork.
Core High-Performance FastAPI & Async Python Capabilities
- Non-Blocking Event-Loop Hygiene: Auditing `async def` endpoints to eliminate event-loop-blocking synchronous I/O or CPU-bound calls (`run_in_threadpool` / ProcessPoolExecutor / Celery), keeping p99 latency under 15ms.
- SQLAlchemy 2.0 Async & `asyncpg` Connection Pooling: Architecting clean repository/service patterns, Alembic migrations, `selectinload` eager loading, and `pgvector` cosine similarity queries.
- Server-Sent Events (SSE) & WebSocket LLM Token Streaming: Streaming structured JSON deltas, tool-call status updates, and citation chunks to React/Next.js frontends with proper client-disconnect cancellation (`request.is_disconnected()`).
Frequently Asked Questions
Why is our FastAPI server hanging under 50 concurrent requests?
The #1 cause of FastAPI latency spikes is calling a synchronous library (such as `requests`, synchronous `boto3`, or synchronous SQLAlchemy/psycopg2) inside an `async def` route—which freezes the entire Python `uvloop` event loop. Switching to `httpx.AsyncClient` and `asyncpg` unblocks concurrency immediately.
How can we hire you as a freelance FastAPI programmer?
Message me on Upwork.