Ali Cagatay

Leading the company’s applied AI research and engineering - taking bold ideas from first experiments all the way to real products in the hands of the end users, and making sure those products are built on rock-solid foundations that stand up in production.

Built the software and machine-learning systems the company runs on day to day - internal products, data pipelines, and the models behind them.

Researched and built machine-learning models for the company’s core products and innovations, and engineered the data pipelines that fed them.

I lead and co-lead Saturday classes on the Intro to Programming course, teaching JavaScript, Python, Git, Jest and SQL to career-changers with no coding background, and reviewing their code in class and through the week.

Built and shipped an Android game, ‘Predict The Number’, to the Play Store in under a month.

A multi-agent RAG over 484 YouTube transcripts from 18 creators - it routes each question to the right creator’s persona and answers in their voice, grounded in what they’ve actually said.

A voice AI assistant for Birmingham City Council that answers housing questions out loud - a Flutter app over an Azure OpenAI and Whisper backend, built in 5 hours with a 5-person team and placed 3rd of 15 at the Microsoft hackathon.

A fast, minimal CRM that helps small businesses track the products they’re selling and who they’re selling them to.

A clean, no-frills Pomodoro timer for timing work and break sessions.

I did my MSc to go deeper into AI after the field really clicked for me, spending the year on machine learning, deep learning, applied AI, data visualisation and the wider impact of AI. The professional-placement part put me inside ProBuild360 as an AI engineer, doing the real job alongside the coursework. For my dissertation I built a multimodal deep-learning model that predicts property prices across the Birmingham housing market by fusing geospatial and temporal data. It’s where most of my deep-learning and applied-AI foundations come from, and I finished with a Distinction.

My BSc came out of loving to build things since I was a kid, and it gave me my core software-engineering foundations - data structures and algorithms, software engineering, and mobile and ubiquitous computing. It was also my first real taste of AI, through modules in machine learning, neural computation and computer vision, and projects like a computer-vision model that classifies brain scans for tumour detection. My final-year project brought the two together: a mobile app that uses NLP and neural networks to recommend workouts based on the muscle you want to train. I graduated with a 2:2 and an Honours.