Asha Rao
AI Engineer | LLM, RAG & Agentic Systems
Summary
AI Engineer specializing in large language models, retrieval-augmented generation, and production agentic systems. Built AI features serving 3M+ monthly users, reducing support costs by 40% and improving answer accuracy from 71% to 93% through systematic evaluation.
Skills & Tools
LLM & GenAI: GPT-4, Claude, LangChain, LlamaIndex, RAG, Fine-tuning, PEFT
ML & Data: PyTorch, Transformers, scikit-learn, Pandas, pgvector, Pinecone
Engineering: Python, TypeScript, FastAPI, AWS Bedrock, Docker, CI/CD
Experience
AI Engineer at Lumina AI
Jun 2022 to PresentSan Jose, CA
- Built an enterprise RAG assistant (LangChain + GPT-4 + pgvector) answering 100k queries/month with 93% grounded accuracy.
- Designed a streaming agentic pipeline with tool calling and guardrails, cutting human handoff rate by 60%.
- Created an eval harness with 2,000+ golden Q&A pairs, lifting retrieval recall from 71% to 91%.
- Reduced inference cost per query by 45% via prompt caching, quantization, and model routing.
Machine Learning Engineer at DataPulse Corp
Jan 2020 to May 2022Palo Alto, CA
- Deployed a fraud-detection model on 2M transactions/day, catching 3x more fraud with 60% fewer false positives.
- Automated feature pipelines with Airflow and Feast, reducing model refresh time from days to hours.
AI & LLM Projects
DocSage | Document Q&A System | Creator | github.com/asharao/docsage
Mar 2023 to PresentOpen-source RAG stack that answers questions over PDFs and web sources with citations.
- 6,800 GitHub stars; 12k monthly demo users; 90% answer accuracy on benchmarks.
AgentBench | LLM Agent Evaluation | Author | github.com/asharao/agentbench
Sep 2023 to PresentEvaluation framework for measuring tool-use and multi-step reasoning in agents.
- Used by 30+ companies; adopted in 2 peer-reviewed evaluation studies.
Education
Master of Science in Computer Science (ML Specialization) at Georgia Institute of Technology (GPA: 3.9/4.0)
Aug 2018 to Dec 2019Atlanta, GA
Certifications
AWS Certified Machine Learning – Specialty | Amazon Web Services (Apr 2023)
Deep Learning Specialization | DeepLearning.AI (Nov 2022)