I build systems that run without me — agentic pipelines, internal platforms, autonomous tooling. Three years in, mostly from zero. Currently in MoEngage's CTO Office.
Seven systems in production, most of them built from nothing — agentic pipelines, internal platforms, the observability underneath them. I own them from the architecture argument to whatever they do afterwards. The three below are the ones worth reading about.
Network simulation and analytics for Verizon — a gNB digital twin fed by live network data, a forecasting engine, and the field-ops portal that put both in front of engineers.
You can't hardcode your way out of every failure mode, and trying to only moves the brittleness somewhere you can't see it. I design for the uncertainty instead — guardrails, retries, and a human with the authority to overrule the machine.
Support IQ scores every ticket, and every score can be disputed by the person it was written about. Automation that people can't argue with doesn't get used twice.
When the P1 bot can't answer, it doesn't return "I don't know" — it returns the trace it already assembled while trying. What a system does when it fails is most of what it's worth.
The harder half is making it runnable by someone who isn't you. I onboarded a second engineer to Support IQ mid-flight, and I still maintain everything I've shipped rather than handing it over and moving on to the next thing.
What I'm looking for: a founding engineer role, or full-stack AI engineering at a startup. Somewhere the AI work is the product rather than a feature bolted on at the end, and where I own systems the way I've owned them for the last two years — from the architecture argument through to whatever they do in production. Email is the fastest way to reach me.