Ships products, not pilots.
I don't just do AI. I've run the ops. I know where the fires are — and I build the tools that put them out. Live apps, real users, shipped from an operator's seat.
Real apps. Real users.
Running in production.
Not demos or decks — shipped software, most of it on Cloudflare, each one solving a job an operator actually has.
ApplyVerified
A job-search operating system plus a vetted-applier marketplace — the full stack from search to a human who applies for you, built on Cloudflare.
LiveIncome Accelerator Coach
A full AI career coach: a phased program with WhatsApp onboarding, a call queue, receipt AI, mock interviews, a skill tree, quests and certificates. Google Ads live, Roman-Urdu mode, partner program.
LiveThe Salary Sheet
A $1 funnel for Amazon-course graduates, with a live wallet payment rail wired end to end.
LiveLinkedAscend
A LinkedIn growth SaaS running on Vercel with Inngest handling the background jobs.
MunchHub POS
An offline Windows point-of-sale paired with a cloud analytics suite — built to keep working when the connection doesn't.
Chanjing toolkit
Digital-human video and text-to-speech, exposed as an MCP server plus REST scripts so other agents and pipelines can call it directly.
Have one that needs shipping?
The pattern above repeats: scope it, build it, put it in front of real users. Fast.
Talk through it →How an ops operator ships AI products
The leverage isn't a single app — it's a reusable toolchain that turns knowledge, process and jobs-to-be-done into shippable software.
Head of AI at Chai Vision
Before the solo products, the discipline — running an AI function with real uptime, real governance and a real P&L.
Led a cross-functional AI Guild — data scientists, ML engineers and PMs — shipping models at 99.9% uptime.
Models moved to production through Git-native CI/CD, with Responsible-AI practice embedded from day one: bias scans, data lineage and model cards as part of the pipeline, not an afterthought.
99.9% uptime via Git-native CI/CD
Shipping ML models as versioned, reviewed, deployable artifacts.
Responsible-AI, embedded
Bias scans, lineage tracking and model cards built into the delivery flow.
Owned the AI P&L
Accountable for the function's cost and value, not just its models.
AI Fluency program + quarterly hack weeks
Raised the whole org's capability and kept a steady cadence of new builds.
Let's ship something real.
Operator who builds, builder who's run the ops. If you've got a product that needs to exist, start here.