Deep dives on problems from my own projects: the trade-offs I made, the code that does the work, and where each approach stops working. For how these choices add up, see How I work.
Why a rate limiter shared by many instances needs the read, the decision and the write to be one indivisible step, and how two short Lua scripts get there at 1.33 ms p95.
Putting an LLM on top of multi-tenant admissions data. How AdmitDesk keeps the agent inside one college and one applicant, and why taking capabilities away beats asking the model to behave.
How Musafir decides that two different trips drive the same stretch of road: slicing routes, sampling points, checking headings, and why the geometry and geography types both show up.