Dhruv Kashyap

A FastAPI backend enabling semantic search and AI-powered Q&A over uploaded PDFs using a Retrieval-Augmented Generation (RAG) pipeline integrated with OpenAI. Features pgvector for vector similarity search and Redis/DiceDB caching to significantly reduce response latency.

• — PostgreSQL schema with pgvector for document embeddings
• — End-to-end PDF extraction, chunking & embedding pipeline
• — JWT authentication + fastapi-guard security layer

An end-to-end ETL pipeline ingesting 6 raw CSV datasets into a Snowflake star schema with PySpark batch jobs for large-scale yield aggregation, rolling averages, and profit calculations. Enterprise RBAC with column-level views.

• — 4 dimension tables, 1 fact table with custom profitability metrics
• — Streamlit & Plotly dashboard with 5 analytical views
• — RBAC: ETL & Analyst roles with secure column-level access

A Spring Boot REST API with full CRUD over MySQL, structured across model/repository/service layers following SOLID principles. Containerised with Docker Compose for single-command deployment and GitHub Actions CI/CD for every push.

• — Automated CI/CD pipeline catching integration issues early
• — All 8 endpoints documented in Postman collection
• — OOP & SOLID design throughout service layer

A FastAPI REST API with 7 endpoints for full CRUD on expenses, category-based filtering, and dynamic total calculation. Input validation enforced via Pydantic models with self-documentation through Swagger UI.

• — Pydantic models for strict input validation & data integrity
• — Proper HTTP status codes & structured error handling
• — Auto-documented via FastAPI Swagger UI + Postman verified