How I think about building software.
Original technical writing by Sanjai Syamaprasad on architecture, AI application development, debugging, system boundaries, and practical product engineering.
How RentNinja turns screening work into a structured application.
Organization-scoped data, authentication, applicant state, AI-assisted intake, and why the workflow model matters more than any single feature.
Five engineering principles I use when a system has to stay understandable.
How I think about boundaries, state, observability, failure modes, and keeping complexity proportional to the problem.
Designing AI features as part of a real application—not as a demo.
A practical approach to model boundaries, validation, fallbacks, user review, and the application state around an AI feature.
Build the workflow first. Add AI where it earns its place.
A framework for deciding where AI belongs inside an application—and where deterministic software should stay in control.
Code shows implementation. Writing shows reasoning.
A repository can show what exists. Technical writing can explain why a boundary, workflow, or integration was designed that way.
These notes are tied to real engineering work and are written to be useful on their own—not as keyword pages. That makes the site stronger for both readers and search engines.