Hands-on with AI
I build daily with Claude, Gemini, ChatGPT, and Copilot — and I'm deep in learning agentic AI and generative AI systems: multi-step agents, tool use, and applied LLM workflows.
My daily driver for backend work. I pair with it on API design, code review, refactors, and debugging — and use its tool-calling to build agentic scripts that read logs, hit internal endpoints, and summarize incidents.
Used for long-context work — feeding entire codebases or spec docs in one pass to trace how a change ripples across services, plus quick research and Google Workspace-adjacent tasks.
My prototyping playground — sketching out GenAI feature ideas, iterating on prompt structures, and building small custom GPTs to test agent behaviors before porting the logic into production code.
Lives in my editor for the daily grind — boilerplate, test scaffolding, and repetitive backend CRUD so I can spend more time on architecture and the trickier logic.
Multi-agent orchestration, tool-calling, RAG pipelines, and evaluation — turning LLMs into reliable backend systems.
Claude-based agent that classifies and routes incoming tickets with tool-calling.
Retrieval pipeline for engineering runbooks, embeddings, and reranking.
Speeding up backend test coverage using in-editor AI pairing.
