Looking for a job

Rajesh Racha

@rajeshracha

Message

AI/ML Engineer at UnitedHealth Group building a production RAG application across 120K internal healthcare documents.

United States

Message

What I'm looking for

I'm looking for AI engineering roles where the team is shipping real products, not just demos.

At UnitedHealth Group, I build and maintain a production RAG application across internal healthcare documents. I work across ingestion, retrieval, model integration, evaluation, and deployment, and reduced LLM token spend by about 40% without measurable quality degradation.

I also develop FastAPI services and AI workflows with grounded citations, structured outputs, and explicit refusal behavior. I built an evaluation set of about 500 questions and use quality gates to prevent regressions from reaching production.

At TCS (Client: Hilton), I built AI and NLP applications for guest-feedback analysis, document classification, semantic search, entity extraction, and an internal assistant over operational SOPs. Fine-tuned Transformer models improved guest-feedback categorization F1 from about 0.71 to 0.89.

At Goldman Sachs, I built production machine-learning systems for fraud detection and transaction-risk scoring. I improved precision in the highest-risk segment by about 22% over the previous heuristic baseline and identified and corrected a temporal data-leakage issue.

Experience

Work history, roles, and key accomplishments

Current

Forward Deployed Engineer (FDE)

Current

UnitedHealth Group

Feb 2025 - Present (1 year 8 months)

Build and maintain a production RAG application across ~120K internal healthcare documents using Python, LangChain, FAISS, FastAPI, embeddings, and Azure OpenAI. Own technical capabilities spanning data ingestion, retrieval, model integration, backend APIs, evaluation, observability, deployment, and production troubleshooting.

Python Lang Chain Faiss fastAPI Azure OpenAI RAG Embeddings MCP Docker Kubernetes Azure DevOps CI CD Langfuse Langsmith Prometheus Grafana TypeScript React Next.js QLoRA LLaMA 2 GPT 4 Claude Amazon Bedrock

AI / ML Engineer

TCS

Jun 2022 - Aug 2024 (2 years 2 months)

Worked directly with Hilton business and technical stakeholders to understand operational workflows and identify AI opportunities. Built AI/NLP applications for guest-feedback analysis, document classification, semantic search, entity extraction, and an internal virtual assistant over operational SOPs.

Python fastAPI SQL PySpark Apache Spark Databricks Kafka Hugging Face Transformers Azure ML MLFlow Power BI Tableau Docker Kubernetes Jenkins Azure DevOps Terraform CI CD

Data Scientist / ML Engineer

Goldman Sachs

Nov 2020 - May 2022 (1 year 6 months)

Built production ML and analytical systems using Python, SQL, XGBoost, logistic regression, and scikit-learn for fraud detection and transaction-risk scoring. Processed 100M+ transaction records using Apache Spark and PySpark, building scalable preparation, feature-engineering, validation, and scoring workflows.