I built an AI CI/CD Failure Investigator & Remediator that uses LangGraph to inspect CI logs and repository code, with hybrid vector RAG to match failures to historical incidents. It adds human approval before opening GitHub pull requests and streams agent milestones through FastAPI.
For my GitOps E-Commerce Platform, I built a MERN application with RESTful APIs and JWT authentication, then automated its delivery with Jenkins, Docker, and SonarQube. I also configured Kubernetes deployment through ArgoCD, including autoscaling and health probes.
During the IBM SkillsBuild virtual internship, I trained and tuned a Stacking Regressor for food delivery time prediction, achieving a 3.16-minute MAE. I developed a FastAPI inference service tracked with MLflow and automated CI/CD to AWS with GitHub Actions; the model reduced RMSE from the baseline IBM AutoAI pipeline's 4.90 to 3.72.
