This website is my public professional and project archive. It collects writing, talks, papers, open source work, and selected project notes that are safe to publish outside my private knowledge base. The site is intentionally simple. The content is the main surface: public engineering themes, open source projects, reliability and operations writing, and posts that connect software systems thinking with physical computing projects such as reef-pi. The current refresh uses a public-only publishing pack generated from my personal LLM wiki.

Looking back across my public talks and writing, the same idea keeps returning in different forms: operational work should become software. In the early 2010s, that meant Chef, infrastructure as code, and testing cookbooks. At PagerDuty, it meant CI/CD for infrastructure, TDD in operations, and service discovery with Consul. At Uber scale, it meant production engineering, capacity management, incident taxonomy, and ML-assisted operational decisions. The tools changed. The direction did not.

reef-pi started as an open source Raspberry Pi reef tank controller. The public record now shows something larger: a hobby controller that grew into a small ecosystem of software, documentation, community support, and compatible hardware. That arc is interesting to me because reef-pi is not only about aquarium automation. It is a compact example of how public engineering systems mature when they need to survive real users, real hardware, and long-lived maintenance.

Reliability engineering often starts with tooling, but the durable work starts earlier: naming what actually fails. My public talks on resiliency and incident management span from DevopsDays India in 2013 to Uber-focused talks in 2020. The shape changed over time. The early framing was culture, tools, and practices. The later framing was incident taxonomy, data, and large-scale operational learning. The through-line is the same: resilient systems are built deliberately, not wished into existence.