SOFTWARE ENGINEER · APPLIED AI · UVA
Prithvi
Raj.
I build AI systems that ship — LLM pipelines, agentic workflows, and the infrastructure to run them reliably.
EXPERIENCE
- —Built 4 prod Django microservices with GraphQL APIs on GKE, serving 1,000+ DoD users with automated failover.
- —Engineered an agentic query builder using LangGraph that translates natural language into complex database searches, reducing query construction time from 90s to 20s.
- —Architected an automated pipeline using Google Cloud Vision and Gemini 2.5 Flash to read and extract data from 200+ vehicle label images per run, cutting manual data entry by ~90%.
- —Reduced data-entry steps by 50% by building 5 React components with centralized state management.
PROJECTS
Argus
Python · Go · Next.js · FastAPI · PostgreSQL · AWS · Terraform
A tool that catches silent LLM behavior shifts — model updates that break apps without triggering a single error — by detecting drift and alerting teams via Slack and a live dashboard.
VIEW PROJECT ↗GitGuard
LangGraph · Claude Sonnet · FastAPI · PostgreSQL · Docker · AWS
A multi-agent system that autonomously reviews GitHub repositories for security issues, performance problems, and documentation gaps — cutting manual review time by ~30 minutes per repo.
VIEW PROJECT ↗FeatherDB
Java · Gradle · JUnit
A lightweight relational database engine built from the ground up in Java, with a custom SQL-like query language, recursive descent parser, and B-tree-indexed storage.
GITHUB ↗SKILLS