AI Query Engine

An enterprise-style AI-powered natural language query engine built with FastAPI, React, OpenAI, Aerospike, and Redis.

Tech Stack

Backend

• FastAPI
• Aerospike (primary NoSQL store)
• Redis (sorted-set cache for top-N queries)
• OpenAI API
• Python

Frontend

• React
• Axios
• Vite

Features

• Natural language to structured query translation (AI Query Planner)
• Deterministic, schema-validated query execution (Pydantic guardrails)
• Hybrid retrieval: Aerospike secondary indexes + in-memory fuzzy search
• Redis-backed O(log n) cache for unfiltered top-N salary/date queries
• Aggregation (MAX, MIN, AVG, SUM, COUNT), sorting, filtering, pagination
• Full observability: execution strategy + per-stage timing metrics
• React frontend with live cache-hit indicator
• One-click startup via BAT scripts

Project Structure

• backend/ FastAPI app, Aerospike + Redis integration, seeding scripts
• frontend/ React UI (Vite)
• scripts/ One-click startup scripts
• docs/ Architecture notes and technical writeup

Run

• Backend scripts/start_backend.bat
• Frontend scripts/start_frontend.bat
• Entire Project scripts/start_project.bat

Architecture

• See docs/architecture.md for the data flow diagram and docs/technical-report.md for the full engineering writeup.

Future Work

• Semantic vector search
• Authentication
• Docker Compose for full-stack orchestration
• Kubernetes
• CI/CD
• Cloud deployment (Vercel / Azure)