Custom Enterprise RAG System with LlamaIndex & Hybrid Vector Search
Project Description & Deliverables
Job Summary
Seeking an experienced AI Engineer to develop an internal knowledge retrieval engine (RAG) over 50,000+ proprietary PDF manuals, contracts, and technical documents.
Key Scope
- Chunking optimization (semantic chunking + parent-child retrieval).
- Hybrid Search implementation combining dense embeddings (text-embedding-3-large) with sparse BM25 indexing (Qdrant/Pinecone).
- Re-ranking pipeline using Cohere Rerank v3.
- Citation verification to guarantee 0% hallucination on financial numbers.
Escrow funded on contract start. Release strictly on milestone verification.
Submit a Proposal
Your proposal will be reviewed by the client. Funds are released based on agreed milestone deliveries.
About the Client
Verified EmployerMarcus Vance
@marcus_vance_2053 • Enterprise Client
Verification Status
- Payment Method Verified
- Identity & KYC Confirmed
- Email Address Verified
- Milestone Release: Fast (< 24h)
Client Summary
Verified enterprise client representing Vance Global Media. Specializing in technology milestones with 100% escrow protection.
100% Escrow Protection
Milestone payouts are held safely in RBI-regulated Khaliwork Escrow and only released when you deliver and the client approves.
Popular & High-Demand Projects
Most viewed client briefs with active proposal submissions.
Fine-Tune LLaMA-3 70B on Custom Legal & Regulatory Datasets (LoRA/QLoRA)
Automated Data Extraction Pipeline with Cloud Scheduler & Postgres
High-ROI Google Search Ads & Meta Performance Marketing Campaign
Full-Stack Next.js 15 App Router & PostgreSQL Web Application
Recently Viewed & Related Jobs
Projects matching your recent browsing history and category interests.