✧ 😎 About Me ✧

Hi 👋, Nice to meet you!

I'm Philip Abakah — AI Engineer, Optometrist, and builder of real-world systems.

I didn't come into this from computer science in the traditional way. I started early — building mobile applications in high school, long before I formally stepped into AI. That foundation in software engineering shaped how I think about building systems: practical, user-focused, and designed to actually be used.

I later studied optometry 👨🏾‍⚕️ and vision science — perception, how people interact with the world — and somewhere along the way, I started asking a different question:

What if we could build systems that don't just see… but actually help people navigate reality?

So I kept building, just in a different direction.

The Work

My work sits at the intersection of AI, healthcare, and accessibility — not as ideas, but as systems that run in the real world.

One of the most defining projects I've built is Safe Step, my final year thesis — an AI-powered navigation system for visually impaired users, built around a real-time computer vision pipeline using YOLOv8. But the model was never the point. The real challenge was translating perception into action — turning noisy visual data into simple, usable guidance. Under simulated blindness, users went from 0% task completion to 100% with the system.

I've also built EyeDxAi, an AI-assisted diagnostic tool that combines image understanding with structured reasoning to suggest possible eye conditions and next steps — pushing me to think about how models can support real decision-making, not just predictions.

The Turning Point

Joining Envision Technologies late 2024 changed everything. Working on systems used by over 100,000 blind and low vision users shifted how I think about building software. It wasn't about prototypes anymore. It was about reliability, usability, and designing systems that people depend on.

It was the first time I saw something I worked on exist outside of me — in the hands of real users, in real environments, solving real problems.

Outside of my core work, I enjoy following Formula One 🏎️ and occasionally exploring telemetry data to build small models. I also enjoy illustration, painting & cooking. See what I'm doing now.