AI/ML • Generative AI • Agentic AI • LLM Engineering
I build AI systems that are measured before they ship.
AI/ML and GenAI engineer in Houston, Texas. I build RAG assistants, agent workflows and MCP servers for a federal agency and a credit union, and I measure the RAG systems with RAGAS.
View resume LinkedIn GitHub Contact
Measured with RAGAS. Bars show each value as a share of its scale.
Hallucinations fell 43% in the same VA work.
I build enterprise LLM applications, multi-agent workflows, MCP-enabled tools, and RAG and GraphRAG pipelines. Each one comes with evaluation, guardrails and monitoring.
My backends are Python and FastAPI with CI/CD and testing. I care about grounded answers, secure tool access and AI quality I can measure.
I raised RAG answer faithfulness from 0.57 to 0.91 on RAGAS and cut hallucinations by 43%, under federal security requirements.
I traced and evaluated runs with Langfuse and LangSmith, and added CloudWatch and Prometheus monitoring with drift alerts.
A RAG and GraphRAG assistant over about 2,000 documents, built on pgvector, Pinecone, ChromaDB and reranking. 25 people across 3 teams use it for about 150 queries a day, which saves roughly 125 hours a month.
Accuracy rose from 75% to 88% and response time fell from 8 to 4 seconds. Prometheus and Grafana track it.
FastMCP servers that 5+ applications use. They add PII and DLP guardrails, audit logging, and OAuth 2.0 with IAM least privilege.
LangGraph and Google ADK workflows with tool calling and human-in-the-loop approval. I gathered requirements from business teams, demoed prototypes to non-technical users, and built their feedback into production.
Details are kept general. These systems belong to my employers and are not public. The projects below are public code.
An enterprise knowledge assistant built on MCP, with controlled tool access, retrieval and agent workflows.
A multi-agent finance system covering portfolio, budget, expense, risk and reporting workflows.
A production-minded LLM application that shows reusable patterns for enterprise AI integration.
More LangChain, Hugging Face, Pydantic, testing and GenAI projects.
LangGraph, LangChain, Google ADK, MCP and FastMCP, multi-agent systems, tool calling, prompt engineering. Models from OpenAI, Anthropic, Gemini and Bedrock.
RAG, GraphRAG, embeddings, hybrid retrieval, reranking, pgvector, Pinecone, ChromaDB, RAGAS, Langfuse, LangSmith, hallucination detection, guardrails.
Python, FastAPI, Pydantic, PyTorch, TensorFlow, scikit-learn, XGBoost, Hugging Face Transformers, SQL, PySpark.
AWS (SageMaker, Bedrock, Lambda, ECS, EKS), Azure OpenAI, Vertex AI, Docker, Kubernetes, Jenkins, GitHub Actions, MLflow, OpenTelemetry, Prometheus, Grafana. I use Claude Code daily.
Senior AI/ML Engineer. RAG and GraphRAG, LangGraph and Google ADK agent workflows, MCP servers, and evaluation and monitoring.
ML Engineer and AI Engineer. LangChain and OpenAI RAG and agent applications, LLM evaluation, and NLP pipelines on AWS SageMaker, Lambda, ECS and EKS.
Data Scientist, after a year as a Data Science Intern. ML models with PyTorch, TensorFlow, scikit-learn and XGBoost, served through FastAPI, tracked in MLflow and deployed on AWS SageMaker.
M.S. Computer Science, Texas A&M University-Kingsville, 2022 (4.0 GPA). B.Tech Computer Science and Engineering, Vignan's University, 2020.
Claude API Fundamentals, Agent Skills, Claude Code, MCP Advanced Topics, and the Deep Learning Specialization.
Send me the use case. I will walk you through my RAGAS evaluation setup. I am in Houston, Texas and open to relocation anywhere in the U.S.
Email me LinkedIn Download resume