I build AI systems that go beyond model demos: retrieval pipelines, graph-based reasoning, cloud-native services, deployment workflows, and the supporting system design behind them.
My current work focuses on GraphRAG, AI system design, cloud deployment on Google Cloud Platform, and practical AI-assisted engineering workflows for real-world development.
Three areas of AI systems engineering I’m currently deepening through projects, system design, and implementation work.
Designing RAG systems as modular, production-oriented architectures — separating responsibilities across services, defining clear API boundaries, and deploying on cloud infrastructure with CI/CD, observability, and operational resilience in mind.
Designing agentic workflows where specialized components coordinate across task boundaries, supported by evaluation pipelines, retrieval benchmarking, and structured memory or context management to keep quality measurable and improvable.
Defining trust boundaries inside AI systems — controlling tool access, validating information before it crosses service boundaries, and adding human or policy-based gates where autonomous actions carry higher risk.
A focused set of repositories that reflect my current AI engineering direction, from flagship system building to workflow and architecture documentation.
My flagship AI engineering project: a GraphRAG-based financial assistant that combines SEC data ingestion, graph construction, retrieval, and cloud service deployment.
Practical shell configurations, workflow helpers, and small operating-system-level tools for daily productivity across macOS, Windows, and Linux.
ML/AI engineering is a systems discipline spanning data preparation, retrieval and storage design, model interaction, backend services, evaluation, deployment, and iteration. This section reflects my current field-level understanding of how those layers connect.