Generative AI Engineer · Portfolio

RushiKedhar.

I build production GenAI systems — RAG pipelines, LangGraph multi-agent orchestration, and harness-engineered LLM evals that ship to real users at Philips Healthcare and the University of Houston.

Tap for sound

About

I ship production GenAI — not demos.

MS Computer Science student at the University of Houston (GPA 4.0), Generative AI Developer & Teaching Assistant, and ex-Philips Healthcare intern. I build agent-harnessed RAG systems, evaluation platforms, and production automations with measurable cost, reliability, and latency outcomes.

70%
grading-time reduction at University of Houston
$800/mo
OpenAI spend saved at Philips Healthcare
0.91
RAGAS faithfulness on DriftWatch evals
4.0
MS Computer Science GPA

Experience

Applied AI work across classrooms, healthcare, and production systems.

Full experience

Aug 2025 — Present

Generative AI Developer & Teaching Assistant

University of Houston · United States

Deployed LLM-powered academic tooling and mentored students on RAG, LangGraph agent patterns, and observable AI runtimes.

  • Built a Python/FastAPI grading pipeline for 100+ students, cutting grading time 70% while keeping hallucinations below 2% through golden-dataset regression tests.
  • Engineered prompt-chaining, verification loops, and scope control that reduced average per-student API cost to $0.003.

Jul 2024 — Jun 2025

Salesforce Developer & Generative AI Intern

Philips Healthcare · Bengaluru, India

Shipped production GenAI workflows inside Salesforce CRM for healthcare documentation and order automation.

  • Built a LangChain/OpenAI/FAISS chatbot with semantic caching and token-budget constraints, reducing documentation retrieval time 60%.
  • Cut OpenAI API spend by roughly 40%, saving about $800/month at production scale.