Designed and implemented agentic AI architectures with tool calling and orchestration frameworks, integrating agents into developer workflows and enterprise tools. Built solutions using MCP (Model Context Protocol) to connect AI agents with developer tools and CI/CD workflows. Created automated evaluation and benchmarking workflows to measure agent output quality and regression risk before deployment. Implemented secure integrations with access controls, approval workflows, and audit trails for traceability and compliance. Containerized and deployed agent services using Docker and Kubernetes; monitored health and reliability with ELK Stack and Prometheus/Grafana.

Python Developer / DevOps-Integrated AI Engineer at Johnson and Johnson

January 1, 2024 - February 1, 2025

Built CI/CD pipelines using OpenShift Pipelines and Tekton to automate testing, validation, and deployment of AI agent updates. Developed retrieval-augmented agent workflows using LangGraph for multi-step tool use and long-running tasks. Deployed agent services as REST APIs using FastAPI and Docker, integrating with existing toolchains. Added automated test and validation steps into deployment pipelines to catch regressions pre-release. Supported reproducible, auditable agent experiments by integrating Git/version control with CI/CD; iterated on prototypes with engineering teams.

Data Scientist / Full Stack Python Developer at American Express

August 1, 2021 - August 1, 2023

Built automated ML pipelines using Kubeflow, MLflow, and Airflow, including test, validation, and rollback steps for safer deployments. Implemented data governance practices with role-based access controls and audit trails. Created benchmark and evaluation suites to measure model/pipeline reliability across versions. Applied explainability and validation techniques (SHAP, LIME) to support reviewable and auditable automated decisions. Collaborated with engineering teams to translate requirements into automation and iterated based on feedback.