Applied-AI Builder
ODAY BRAHEM
I build AI products end to end and run them solo; three are live right now. When the tooling I need doesn’t exist, I build it, and lately I’ve been open-sourcing it.
I also write nextbig.dev, a daily briefing on AI infrastructure.
Work
-
A voice assistant that sees your screen. Hold a key and ask: it answers out loud, dictates into any app, and clicks and types to do the task itself. Speech-to-text and the voice run on the machine, and the mind is the ChatGPT plan you already pay for.
-
A local cost meter and spend cap for AI coding agents. It prices usage from Claude Code and other agents into a local ledger, serves a dashboard on localhost, and stops the spend at the cap you set. One static binary, and everything stays on your machine.
-
A daily briefing on AI infrastructure. An LLM pipeline reads 300+ sources and publishes the briefing, a podcast in a synthetic version of my voice, and one falsifiable call a day, settled in public. Agents can pull all of it over MCP.
Today“Fable 5 is back at full capacity, and Anthropic’s answer to a brutal week is a model you can actually call”
-
Online voting for HOAs and communities. Ballots and elections, support agents that turn tickets into bug reports before I see them, and a published, state-by-state corpus of HOA voting law.
-
Autonomous drone flight, trained vision-only on rented cloud compute. I started from scratch and it cleared Anduril’s AI Grand Prix qualifier a couple of months in. Robotic arms on the bench, hardware build in progress.
Why AI
I’ve read obsessively since I was a kid. My mom kept me supplied with science magazines and I read them cover to cover. That curiosity never left: I still want to know everything about anything new, and progress is the part I find most exciting.
AI is the biggest shift I’ve seen since the internet, so going deep came naturally. Lately the focus is infrastructure: what a token costs to serve, who pays for the buildout, and where both are heading. nextbig.dev is that habit, published daily.
History
- 2017Business development at VersaFile, an IBM RPA partner at the time. Got my foot into enterprise software.
- 2018Collabware. Within six months I’d researched the US Navy and Marine Corps’ records system and walked us into the Pentagon with the SharePoint-based proposal the company shipped.
- 2019Moved into engineering. Lead QA engineer for 2.5 years across cloud and on-prem, building test automation to 85% coverage.
- 2021Ran the on-prem-to-cloud customer-success program and cut churn to 4%, then moved into sales engineering and built the demo environments.Alongside it, served as an intelligence operator in the Canadian Armed Forces reserves: OSINT and geospatial, through 2023.
- 2025Back at VersaFile, by then an SAP middleware product company. Migrated a customer base onto SAP S/4HANA and grew revenue on migrated accounts 5×.
- 2026All in on AI, solo: ~6,000 commits in a year across vote.direct, nextbig.dev, OP8, Sonitura, and burnban.