Hire
Hire me by technology
Freelance and consulting, by the stack you need. Every page states what I've actually shipped with it — and what I won't pretend to know.
fastapi
FastAPI is my default for new async APIs and model-serving endpoints — validation with Pydantic, background work with Celery, and webhooks that don't block the request thread.
Hire for fastapiyolov8
I build detection pipelines that keep up in real time on the hardware you actually have: YOLO for detection, DeepSORT/ByteTrack for identity tracking, and export to ONNX/TensorRT when it needs to run on a Jetson at the edge.
Hire for yolov8langchain
I build retrieval-augmented pipelines with LangChain, FAISS and Pinecone over clients' own documents, and conversational agents on LLaMA and OpenAI APIs — with deliberate decisions about what context to carry forward, summarise, or drop.
Hire for langchaincelery
Celery + Redis is how I move slow work — inference, message processing, database writes — off the request thread.
Hire for celeryredis
I use Redis daily: as a Celery broker, for sliding-window deduplication (30-second alert windows on the CCTV build), rate limiting on API backends, and short-lived state.
Hire for redisopencv
OpenCV is the layer under every vision system I build: pulling frames from RTSP/CCTV streams, pre-processing, handing off to detection, and keeping the whole pipeline in real time on the hardware you have.
Hire for opencvtensorrt
When a model has to run on a Jetson instead of a GPU server, I export it through ONNX to TensorRT, benchmark it on your hardware, and report what you actually get — not what the marketing page promises.
Hire for tensorrtpostgresql
Schema design, SQLAlchemy migrations, and the discipline of not breaking a running product: I migrated a client's database from MySQL to Supabase/PostgreSQL mid-project while their AI agents kept running.
Hire for postgresqlnextjs
I'm a backend engineer first, but I've shipped React frontends on real projects — including this site, a static-export Next.
Hire for nextjs