// hello world !! Welcome to my portfolio
I architect systems at the intersection of distributed systems, backend engineering, and machine learning. I build infrastructure that scales to millions of users.
<!-- about.html - Hiten Raju -->
// who I am ยท what I do ยท where I build
Hi! I'm Hiten Raju, a software engineer specialising in backend systems, distributed infrastructure, and AI/LLM systems. Currently an AI Engineer Intern at Keck School of Medicine, USC and pursuing an MS in Computer Science at USC (Aug 2025 โ May 2027), building on 3+ years of industry experience.
// projects.js : things I've built & shipped
const projects = [ ...shipped, ...building ]
Full-stack AI journaling app shipped in 48 hours; RAG pipeline embeds entries as vectors and generates mood-aware responses. Won 1st place among 40+ teams at TechWeek Hackathon. Sub-100ms p99 latency, 30% lift in 7-day retention, 50+ beta users post-hackathon with zero infrastructure changes.
RAG-based screening assistant grounded in resume and JD content; LangChain maintained conversational state, improving answer precision by 35% and cutting hallucination rate by 60% vs. a prompt-only baseline. Reduced p95 latency from 4.2s to 1.1s (74%) via request batching and KV-cache-aware prompting.
Real-time pipeline processing 30 FPS streams at 90%+ accuracy across 8 classes. Adaptive frame sampling cut inference compute by 45%, enabling edge deployment at sub-200ms latency. Presented at IRCICD'23.
// skills.json โ tech stack & tools I actually use
{ "status": "always_learning", "passion": "immeasurable" }
// experience.ts โ professional journey
interface Career extends Timeline {}
Built and shipped multi-agent LLM systems in production, orchestrating autonomous pipelines that extract structured data from unstructured clinical datasets โ reducing manual annotation overhead by 60% while maintaining full HIPAA compliance. Designed agentic AI frameworks with tool-use, persistent memory, and multi-step planning using LangGraph / LangChain. Implemented RAG pipelines over large-scale medical corpora, cutting hallucination rates in safety-critical clinical inference tasks deployed on HIPAA-compliant Azure.
Architected an observability platform adopted by 100+ engineers across 6 teams, reducing mean incident troubleshooting time by 30%. Led migration of 5 monolithic services to a 12-service distributed architecture, cutting deployment time from 30 min to under 2 min (15ร) with zero-downtime rollouts. Redesigned API layer patterns, reducing p95 latency from 30s+ to sub-3s (10ร). Automated CI/CD pipelines across 8 services, eliminating 10+ hours/week of manual effort and reducing deployment errors by 26%. Built SLO monitoring contributing to 99.9% uptime across multinational clients.
Delivered payroll reporting integration for 30K+ employees; independently resolved schema mismatches affecting 15% of records before a hard deadline. Optimized 5 high-traffic database endpoints via indexing and query redesign, cutting peak CPU by 60% and saving $2K/month in cloud costs โ scripts adopted by infrastructure team post-internship.
Built a full-stack invoice workflow platform with RESTful APIs using Spring Boot (Java) and React, processing 1,000+ invoices/month and increasing throughput by 37%. Trained and deployed a payment-risk model on 40K+ transactions using gradient-boosted classifiers, reducing AR team manual triage time by 40%.
/* contact.css โ let's build something */
// open to work, collabs & good conversations
# README.md
๐ USC MS Computer Science (Aug 2025 โ May 2027) ๐ฅ AI Engineer Intern @ Keck School of Medicine, USC (Apr 2026 โ Present) ๐ข Ex-Software Engineer I @ Computacenter (2023โ2025) ๐ค Multi-agent LLM systems cutting clinical annotation overhead by 60% ๐ Led 5-service โ 12-service migration (15ร faster deployments) ๐ TechWeek Hackathon 1st Place (40+ teams) ๐ Research presented at IRCICD'23 Tech Stack: Python ยท TypeScript ยท Java ยท Go ยท LangGraph ยท LangChain ยท RAG Node.js ยท FastAPI ยท Kafka ยท Redis ยท Kubernetes ยท Docker ยท AWS ยท Azure Currently: Looking for full-time SWE roles starting 2027 Let's connect: hraju@usc.edu