Full-Stack/LLM Development (Vietnamese)
Publicada el 2026-07-20
Descripción de la oferta
I’m replacing our summary-based workflow with a true Retrieval-Augmented Generation pipeline and need a collaborator who can own both the backend and the React front end. The core objective is accuracy: every PDF we receive from clients must be chunked, embedded, stored in Qdrant, and then retrieved at query time so the language model can cite exact passages, never a lossy summary. You’ll wire this flow together with Python services and Node/Fastify APIs, expose it to the UI in Typescript/React, and keep everything humming inside our OpenAI-powered evaluation layer. What I’ll lean on you for • Architecting and implementing the end-to-end RAG pipeline (chunking strategy, embedding jobs, vector-store schema, retrieval functions). • Building real-time document retrieval endpoints that push grounded evidence straight into assessments, role-plays, and feedback modules. • Instrumenting detailed logging and audit trails so compliance teams can trace every answer back to source text. • Crafting a clean, responsive React interface for document upload, status monitoring, and citation-rich results. Stack you’ll touch: Python, Node, Fastify, Typescript, React, Qdrant, OpenAI, plus whatever lightweight ops you prefer for deployment. Logistics We’ll start part-time with daily overlap in GMT+8; if we click, there’s plenty of runway to extend the engagement. I’m hands-on and will be building alongside you, so expect tight feedback loops and a true collaboration. Must be available to work 4 hours per day Must be available as soon as possible once you hired If making embeddings trustworthy and auditable gets you excited, let’s talk.
Skills
- Node.js
- Vietnamita
- RAG
- TypeScript
- Python
- MongoDB
- FastAPI
- JavaScript
- Compliance
- React
- Auditoría
- Feedback
- Fastify
- Vector Databases
- SQL
Fuente original: freelancer