What we've actually built.
Real systems, in detail. Pick one to read the full build, from the problem to how it works.
Taking a multi-tenant finance agent into production
An AI accounting platform's agent reads invoices and bank statements, reconciles them, and posts the results into each client's ERP. We built the production system around the model, so it can be trusted with real money for many companies at once.
Multi-tenancy for finance agents, and why row-level security has to be forced
An AI agent reading and writing real invoices for many client companies out of one database. The tenant filter that lives in application code is a promise the model never made, so we moved the boundary into Postgres and made it impossible to merge a table without it.
Automated incident root-cause, straight into Slack
When something breaks at 2am, the Datadog alert is only the start. The real work is figuring out why, fast, while downtime ticks and the on-call engineer plays detective across a dozen dashboards.
An automated local-SEO engine for a taxi-booking business
CityTaxis competes for riders across the cities it serves, which means ranking in local search and on Google Maps. That is a grind that usually eats a marketer's week.