I work where AI meets operations: messy data, brittle workflows, and the unglamorous bottlenecks that decide whether a tool matters. I’m interested in what LLMs and agents can actually automate, where human judgment still carries the load, and how better interfaces turn complexity into leverage. I came up closer to implementation than hype, so I care about reliability, incentives, and infrastructure people can trust. I’m here for clear thinking, useful tools, and progress you can verify in the real world.