Ali
Hamza
Kamboh.
I automate work that used to take a team. AI systems, the backend architecture under them, and the frontend on top.
Names I have worked with
How I got here
My background.
Hey, I’m Ali.
I was building robots and electronics at home before I wrote any software. The code started in 2018, with paid work won from cold messages while I was still at school.
Then ApplyAce, one form so a student could apply to several universities at once, and after that FTMO, where millions of trading accounts were being checked by hand. I built the engine that did it instead and rebuilt it in Rust when it stalled. The review desk went from about 1,023 people to 5.
Since August 2024 I have been CTO at Xaibridge, building for prop firms, healthcare and operations teams on an architecture carrying up to 1M+ users. The question most days is what something costs, and what it would cost if we built it ourselves. I work remote, on site or hybrid, in any time zone.
2018
since
day one
Where I’ve worked
Experience.
- Built end-to-end infrastructure for up to 15 prop-trading firms: MT4/MT5 ingest, real-time metrics, rules engines, payments, trader portals. Also healthcare products, CRMs, admin portals
- Own the architecture: 1M+ users, millisecond-level responses on 3G/4G/5G, agent harnesses, review gates, eval loops
- Cut org token use from ~100B to ~10B tokens/month via routing, caching, quantization and tighter harness design
- Lead engineering, hiring and delivery. Trained 500+ engineers and client teams
What I take on
Services.
Work I take on myself, from the first call to the deploy. Each one says what ships and how long a first version takes. The through line is the same in all of them: build the thing you are currently renting. That is where the saving comes from, in the AI spend and the server bill at once, and again in the vendors charging per user.
Things I’ve built
Projects.
Click any project for the detail
01
Xotion Studio
Creative engine · Node.js
Writing
Some of these builds get written up: how the thing was put together, what broke on the way, and where I stand on the choices it took.
Read the blogAcademic background
Education.
Bachelor of Science, Computer Science
currentUniversity of Management and Technology
Research: optimizer efficiency for LLM training (Muon-family methods)
Results & open source
Achievements.
Trade-evaluation engine taken from 1K to 1M+ checks an hour
Cost per user $5-10 to $0.24
KYC brought in-house off a vendor, ~$3 a verification to ~$0.001
200K+ verifications a day
One org's token use cut from ~100B a month to ~10B
Same output
About $102M taken out of client cost in two years
My own maths at Xaibridge: vendor bills replaced, token spend cut, infrastructure sized down
Review team went from about 1,000 people to 5
Same trade-evaluation work
notebookllama · 63 stars
Open source · github.com/ahkamboh
agent-caption · 22 stars
Open source · github.com/ahkamboh
stepmold, agent-eyes and thinktank
Open source · github.com/ahkamboh
Researching optimizer efficiency for LLM training
Ongoing
Courses & credentials
Certifications.
Let’s connect
Get in touch.
Remote, on site or hybrid. A job or a one-off build, in any time zone, kept low cost and built to scale. If you’re short on time to launch, say what you need and when you need it.
Let’s talk, 30 minutesOr contact me directly