I build software that has to actually work — in clinics, in banks, in the field, offline, on a phone that is five years old.
Software engineer in Pittsburgh. Fifteen-plus years of building products in industries where the requirements are written by regulators and the users are not optional.
No agency menu. These are the things I am genuinely good at. Tap one to see what that actually means.
Here is what the work actually involved — the constraint that shaped it, and what shipped.
A consumer app that teaches long-term investing by letting people simulate it instead of reading about it. Native on both platforms against a Rails API, with projection math, charting, and a content pipeline the in-house team runs without me.
Geospatial capture for crews working where connectivity is not a given. Offline-first sync, map tiles cached to the device, and a reconciliation layer for edits that come back from the field out of order.
Aggregation and reporting across a network of behavioral health providers under real regulatory constraints. The dashboards were the easy part — the work was the ingestion, the data hygiene, and the audit trail underneath them.
Content delivered to people standing in the middle of the thing it describes. Maps, media, and an admin surface that non-technical staff use daily without calling anyone for help.
Phones talking to physical devices over BLE — pairing flows, firmware handshakes, thermal printers, sensors. Including the mock peripherals that let a team test all of it without a lab full of hardware.
Organizations with real stakes and no budget for waste. Tools and sites where a broken form means somebody does not get the help they came for, built to be maintained by whoever comes next.
Years of watching software projects succeed and fail produce a short list of opinions. These are mine.
Most failed projects were mis-scoped, not mis-built. I would rather spend the first week arguing about what we are actually making than spend the last month discovering we disagreed.
Something real in front of users early, then again, then again. A vertical slice that works end to end tells you more in a day than a spec tells you in a month.
Every engagement should end with your team able to run the thing without me. Documented, conventional, boring where boring is correct. Cleverness is a cost you pay later.
Regulated data, bad connectivity, old devices, small budgets. These are not edge cases to handle at the end — they are the shape of the problem, and designing around them from day one is cheaper than retrofitting.
New builds, stalled projects, second opinions, and rescues. If you are not sure whether it is worth doing, that is a fine conversation to start with.