Computer Science × AI × Automation
Humayra Shajahan
Computer Science Undergraduate — AI, Automation & Software Engineering
Final-year Computer Science undergraduate at the University of Westminster, delivered through IIT, Sri Lanka. I build AI-powered tools, automation pipelines and software systems for real operational problems — from market reporting at a stock exchange to enterprise ERP environments.
How the work flows
75 min → <10 min
Manual reporting time
21 weeks
AI upskilling programme
1 year
Enterprise software engineering
5 phases
AI adoption programme
What I actually do
Problem → System → Outcome
Real projects built around operational problems, automation and applied AI.
Automated a daily market reporting pipeline for QSE, reducing a 60–75 minute manual process to under 10 minutes with scheduled PDF delivery.
75 → <10 min
A retrieval and local-LLM assistant that answers market questions without cloud infrastructure.
Query flow
From developer-dependent reporting to self-service analytics.
Report builder — demo
Designed for a non-technical Operations team and anchored to real market-operation use cases — teaching people to work with AI, not just deploying it for them.
This demonstrates that I don’t only build AI systems — I help people adopt them.
Full-stack web app
A platform-style application exploring matching and discovery flows.
Java, OOP
A system for logging, tracking and resolving support tickets.
Java, Database Systems
A database-backed system for managing flight and booking records.
Python
A program modelling and predicting academic progression outcomes.
Web Development
An awareness website built around UN Sustainable Development Goal 14.
If something happens repeatedly, it should probably become a system.
A technically impressive system is useless if the user cannot understand it.
Market Copilot was designed to work locally on an Intel i3 without a GPU.
Qatar Stock Exchange
IFS
Liulian Tech
InkTouched
I also design things for people, not just systems.
Client Name
Live site →
Client Name
Live site →
Client Name
Live site →
Founder · 2021 — Present
What I learned from turning a repetitive market-reporting workflow into an automated system.
Why combining traditional ML with a locally hosted LLM made sense under real hardware constraints.
What changes when AI education is built around real operational problems instead of theory.
University of Westminster / IIT Sri Lanka
2023 — Present
Edexcel International Advanced Level — Mathematics, Physics, Business Studies
G.C.E. Ordinary Level — 7 A’s
Let’s build something useful.