01
Finance & Performance System
A finance and performance system that turned recurring spreadsheet handling into verified reporting across 101 locations.
Hamza Shaikh · Dubai
I build AI systems and useful software for real work.
I take messy problems, work out what is actually happening, then design and build the smallest useful system. That has included finance reporting, websites, operations tools and my own personal AI setup.
101 locationsin a verified finance reporting scope
303 totalschecked against the source
2 live websitesdelivered for real businesses
HarsOSused in my own work every day
Selected work
What I built, and what the evidence proves.
Every case separates a delivered result from an estimate, prototype or next hypothesis.
↗
02
HarsOS
How Hamza Shaikh built HarsOS to connect research, durable context, decisions and useful next actions across real projects.
03
Fortune Travel Operations
I mapped enquiries and ticketing deadlines across WhatsApp, email and Travelport, then designed a shared follow-up board with reminders and approval checks.
About
The work moves between research, operations, product decisions and code.
I’m Hamza. I learn by working on real things, then write down what changed, what failed and what I can honestly prove.
byHARS is the public home for that work. It is also where you can bring me a problem and work through it with me.
Writing
Notes from building and testing.
7 September 2026 · 6 min read
The hard part of finance automation starts after the spreadsheet arrives
I kept the familiar Excel inputs, standardized what happened next, and verified 303 monthly location totals before calling the reporting system finished.7 September 2026 · 5 min read
Why I built a personal AI system instead of collecting more chatbots
The problem was not access to capable AI. It was losing useful decisions, research and unfinished work whenever the tool or session changed.7 September 2026 · 6 min read
What a live AI call coach has to prove before I trust it in a real conversation
The cues became more relevant, but the live test exposed the two failures that mattered most: they arrived too late and the budget could stop them mid-call.Free 30-minute working session
Bring one real problem. Leave with a clearer next move.
No polished brief required. We can compare tools, untangle a workflow, pressure-test an idea or decide that AI is not the useful answer.
Book a session ↗