Email • Portfolio • LinkedIn •
I’m an AI Engineer specializing in production-grade LLM systems, multi-agent orchestration, and scalable RAG pipelines on Microsoft Azure.
✅ 70% latency reduction (8.5s → 2.5s)
✅ 55% cost savings ($12K → $5.4K)
✅ 95% intent routing accuracy for 1,000+ enterprise users
✅ Strong focus on retrieval quality, evaluation, observability, and guardrails
✅ End-to-end delivery from LLM prototypes → full-stack production deployments
📅 Feb 2023 — Present
🧠 LLM Systems | RAG | LangChain | LangGraph | Azure OpenAI | FastAPI | Node.js
- Built an enterprise multi-agent AI platform with LangGraph (HR, IT, Legal, General Query) for 1,000+ employees
- Implemented GPT-4 based intent routing with 95% classification accuracy and 2s P95 latency
- Designed a production hybrid RAG pipeline for 50K+ ServiceNow docs using:
- Dense vectors (text-embedding-3-large, 3072-dim) + BM25
- Azure AI Search with metadata filtering + semantic reranking
- Achieved 0.89 MRR@10 and reduced irrelevant answers 35% → 4%
- Reduced response latency 8.5s → 2.5s using caching + parallel retrieval
- Added observability via LangSmith + Azure App Insights + guardrails (PII + RBAC)
- Reduced monthly LLM costs by 55% with prompt compression + model routing + caching
🚀 Full-stack real-time trading simulator built with microservices
🔗 Repo: https://github.com/VenkataRohan/stock_exchange
Tech: TypeScript, Node.js, React.js, RabbitMQ, PostgreSQL, Prisma, Docker
- 4-microservice architecture with async order processing
- Real-time updates via WebSockets + NGINX load balancing
- Durable processing with retries, acknowledgements & idempotent consumers
✅ LightDash — Threshold-based alerting
🔗 lightdash/lightdash#12119
✅ Rocket.Chat — Notify on Reactions feature
🔗 RocketChat/Rocket.Chat#32475
✅ Cal.com — Booking flow refactor for question sequencing & integrity
🔗 calcom/cal.diy#18192
- Building agentic AI systems using LangGraph
- Improving RAG performance using hybrid search + reranking
- Developing evaluation frameworks for LLM quality + reliability
- Exploring tool calling, workflows, and multi-agent architectures
- 💼 LinkedIn: https://linkedin.com/in/venkata-rohan
- 🌐 Portfolio: https://venkatarohan.vercel.app
- 📧 Email: venkatarohank@gmail.com
- 📱 Phone: +91-94911-11949
⭐ If you're building **LLM products / RAG platforms / agentic workflows / full stack applications **, I’m happy to collaborate.