Computer Science (Artificial Intelligence) student who turns messy data into clear decisions and builds systems that work.
Three gridworld environments in Python, each testing a different AI safety problem: safe interruptibility, avoiding irreversible side effects, and reward gaming. I built agents with different policies and showed how a reward exploiting agent can farm reward without doing the real task.
Automated gate on an ESP32 with a PIR motion sensor, pushbutton override and servo motor. Simulated in Wokwi first, then I wrote the embedded C++ logic and refined it through structured testing and debugging.
Routed networks configured with RIP, DHCP and VLANs in Cisco Packet Tracer, plus an Ubuntu Linux VM set up from scratch. Scored 87% on the assessed coursework.
Bright Network Internship Experience UK, Data and AI track. Validated and analysed 5,000 telemetry readings from offshore wind turbines in Python, flagged turbines past temperature and vibration thresholds, and recommended a real time streaming approach for monitoring.
Four person Java application for a community item sharing library. I built the core class hierarchy, an abstract Item class with Book, DVD and Magazine subclasses, so new item types slot in without changing existing code.
Completed through Reed and Coursera
I'm open to graduate schemes, internships and entry level roles in technology, data and AI, anywhere in the UK.
umehmayac@gmail.com 07506 711580