Headline

AI Engineer | LLM Systems, RAG, Agents (LangGraph, LangChain) | AWS Bedrock | TypeScript, Python, React | Built 0 to 1 AI Healthcare Platform (Ex-Diagna AI)

About

AI Engineer with 2.5 years of experience building production LLM systems end to end: RAG, multi-agent workflows, and the backends and frontends around them. Highlights • Built the document-AI platform at Diagna AI (HIPAA-compliant) that classifies and extracts data from 1,000+ medical documents/day across 40+ categories for live US clinics, using Claude, Amazon Nova and Llama on AWS Bedrock • Cut manual fax triage from about 3 hours to 10 minutes a day and referral turnaround from 4.2 days to same day • Shipped VoiceGenie, a generative-AI voice sales platform, from 0 to $10K MRR in 11 months as the sole frontend engineer What I work on • LLM systems: structured outputs, model routing and fallbacks, grounding, human-in-the-loop review, evals • Agents and RAG: LangGraph, LangChain, hybrid retrieval (FAISS + BM25), RAGAS • Distributed backends: AWS (S3, SQS, SNS, Bedrock), BullMQ/Redis, PostgreSQL, idempotent pipelines • Product frontends: React, Next.js, TypeScript Projects • PayOps AI: 19-node LangGraph multi-agent system for payment reconciliation, with parallel agents, Postgres checkpointing, four-eyes approval and 7/7 golden evals passing • ClinRAG: hybrid RAG over FDA drug labels with a RAGAS eval harness • MotionStudio: browser video editor with in-browser and cloud (Remotion on AWS Lambda) export Stack: Python, TypeScript, Node.js, FastAPI, React, Next.js, AWS Bedrock, LangGraph, LangChain, PostgreSQL, Redis, Docker Open to AI Engineer / GenAI roles. Reach me at ishivamgovindrao@gmail.com

Location

Gurugram

AI stack

claude code

codex

langchain

langgraph

livekit

mcp

Work experience

Software Development Engineer at Diagna, Jan 2025 - Apr 2026 · 1 yr 4 mos, Gurugram, Haryana, India, • Architected and owned the core backend of a HIPAA-compliant clinical document platform serving live paying clinics: 40+ REST endpoints (TypeScript, Express, JWT + RBAC, Swagger) over a 25-model PostgreSQL/Prisma schema • Integrated Claude, Amazon Nova and Llama via AWS Bedrock to classify and extract structured data from 1,000+ medical faxes/day across 40+ document categories, with model fallback, retries and schema-validated JSON output • Cut manual fax triage from about 3 hours to 10 minutes a day and cleared a recurring 2-day backlog by chaining AI extraction, patient matching, human review and EHR write-back • Engineered the distributed AWS ingestion pipeline (S3, SQS, SNS) with 7 BullMQ/Redis queues, DB-level idempotency and a dead-letter queue that self-healed transient failures • Built human-in-the-loop safety: ambiguous patient matches routed to review instead of overwriting records, with transactional writes and AI-extraction lineage • Automated a no-API legacy EHR (MDLand) with a Python RPA service (FastAPI, Robocorp, Playwright, Redis/RQ) plus Medplum FHIR sync, cutting referral turnaround from 4.2 days to same day • Delivered both production React 19 frontends: the FAXFlo clinical inbox with a dual-view AI document editor (Zustand, PDF.js) and an internal patient-record merge dashboard

Software Engineer at Oriserve, Dec 2023 - Jan 2025 · 1 yr 2 mos, Hybrid, • Built VoiceGenie, a generative-AI voice sales platform, from 0 to 1 as the sole frontend engineer (Next.js, React, TypeScript): customer dashboard, AI campaign builder, onboarding and post-call analytics; the product reached $10K MRR within 11 months and closed its first enterprise customers • Built per-campaign ElevenLabs voice controls and an @token script composer that personalized every AI call at runtime • Shipped post-call AI intelligence (transcripts, emotion analysis, lead extraction) synced to HubSpot and GoHighLevel, with Cal.com booking • Integrated Stripe payments end to end (checkout UI, backend APIs, transaction-state handling) • Cut page load time by 30% (LCP, FCP, TBT, CLS) through Lighthouse profiling, code splitting, lazy loading and caching • Collaborated on Twilio voice-AI infrastructure for 2K+ concurrent calls and rate limiting that blocked 15K+ abusive requests/day

Education

Bachelor of Technology - BTech, Computer Science and Engineering, Institution of Engineering and Technology Ayodhya

Public links

github: https://github.com/BLAZE7SHADOW

shivamgovindrao.com: https://www.shivamgovindrao.com/