Build AI Contact Search Platform

I’m building an AI-powered web app that lets people search their own professional network as easily as they search the web. Users will be able to upload contact data coming from CSV files, LinkedIn exports, or Google Contacts; the system then parses and indexes the information so that a natural-language query—typed or spoken—instantly returns the most relevant contacts along with a short “why this matches” explanation.

Core flow
• Secure upload and parsing of the files above
• Extraction of company, industry, location, and skills, then storage in a vector-friendly database
• Natural-language and voice search that ranks contacts semantically and returns short rationale sentences
• Clean, responsive UI (desktop and mobile) that shows results in a clear card or list view with share / copy options

On the tech side I’m leaning toward a React or Next.js front-end, a Node.js/TypeScript or Python backend, OpenAI embeddings (or similar) for semantic search, and a vector store such as Pinecone, Supabase, or Qdrant—but I’m open to your proven stack if it achieves low-latency, accurate results and is production-ready.

Deliverables
1. End-to-end web application deployed to a cloud host (AWS, Vercel, or comparable)
2. Source code in a Git repo with clear README and environment setup scripts
3. API documentation covering upload, search (text and voice), and result schema
4. Basic test suite demonstrating correct parsing, indexing, and retrieval logic
5. Short Loom or written walkthrough showing the system in action with sample data

Acceptance criteria will be a live demo where I upload real contact exports, ask, for example, “Find a fintech founder in Bangalore,” and receive accurate matches with concise explanations referencing company, industry, location, or skills.

If you’ve shipped something similar—search, embeddings, or contact management—let’s talk through your approach and timelines.