Lalitpur, Nepal · Kathmandu University ·
Sandesh builds
full-stack products,
backend to inference.
Final-year Computer Engineering student building full-stack products with Django REST Framework, Next.js and FastAPI, and increasingly wiring machine learning and retrieval into them. This page is built the same way every project is: shipped end to end, not mocked up.
Signal chain
- 01
Frontend & Full-Stack
Channel A
- 02
Backend & APIs
Channel B
- 03
AI / ML & Retrieval
Channel C
- 04
Data & Real-Time Systems
Channel D
Skills, running live
Four modules, each rendering itself instead of describing itself.
Channel A
Frontend & Full-Stack
Product-shaped work, from a Django admin panel to a Next.js storefront, shipped end to end.
- Next.js / React
- TypeScript
- Django REST Framework
- REST API design
- eSewa payment integration
Channel B
Backend & APIs
APIs and auth built to hold up under real traffic, not just a demo.
- Django & DRF
- FastAPI
- Node.js / Express
- WebSockets
- JWT auth
Channel C
AI / ML & Retrieval
Retrieval and multi-agent pipelines that turn raw documents into structured answers.
- TensorFlow
- NLP
- RAG pipelines
- FAISS vector search
- Tesseract OCR
Channel D
Data & Real-Time Systems
The layer under the UI: databases, sockets and protocols moving data reliably.
- PostgreSQL / MongoDB
- Real-time WebSocket streams
- HTTP & MQTT (PyroAPI)
- Adaptive chunk sizing
- SQLite
Selected work
Four projects, four different kinds of hard.
A full-stack online bookstore with real checkout, not just a product grid.
- Authentication, catalog browsing, cart management, checkout and order history built end to end.
- Integrated eSewa test payments for a real Nepali checkout flow.
- Built Django admin workflows for managing books, authors and publications.
About
I'm Sandesh Upadhayay, a final-year Computer Engineering student at Kathmandu University in Lalitpur, Nepal, aiming for software engineering, backend, and AI/ML roles. Most of what I know I picked up by shipping full products rather than isolated exercises.
My stack leans full-stack: Django REST Framework and FastAPI on the backend, Next.js and React on the front end, PostgreSQL or MongoDB underneath. Lately I've been pulling machine learning into that stack too, retrieval pipelines with FAISS, OCR with Tesseract, and multi-agent systems that read documents and reach a consensus.
Outside of coursework I build things with friends when a problem is interesting enough: a bookstore with real payment integration, a marketplace for Nepali handicrafts, a download manager that adapts to network conditions. I'm fluent in Nepali and English, and comfortable working across a repo from admin panel to inference server.