šŸŽÆ
Published on

Back to Coding again

If you’re new here — hi šŸ‘‹, I’m Kevin DEGILA, an AI engineer from Benin Republic in West Africa. I love learning new things, experimenting with code, and teaching what I learn.

Since 2019, I’ve been exploring how AI can empower people. Along the way, I’ve built a YouTube channel with 13,000+ subscribers, published a Deep Learning course with 800+ students on Udemy, and developed several AI demos and prototypes. I’ve worked as a freelancer, in a startup, in a telecom company, and now lead the Data & AI team at the Digital Agency of Benin, where we’re building the country’s AI first use cases.

Two Years Away from Code

For the past two years, I haven’t had a single week of deep, focused coding. I’ve only released five YouTube videos and spent more time managing, leading, and strategizing than writing Python scripts.

And honestly, I don’t regret it. It was a necessary break to focus on impactful national projects, learn how to lead a team, and even discover new joys like playing tennis. But a part of me missed the feeling of solving problems line by line, the creative chaos of building something from scratch.

The Comeback Plan

Now, it’s time to come back to what truly drives me: learning, creating, and teaching.

Over the next few months, I’m diving deep into AI again, not just using models, but understanding them from the inside out. I want to rebuild my foundations, relearn what I’ve forgotten, and push myself technically.

Here’s my plan:

  • Go back to Jupyter notebooks and Python scripts.

  • Follow the ā€œSolve Itā€ method : I enrolled myself in the second cohort of the solveit course by the folks at Fast.ai. More on that on future blog posts

  • Use Andrej Karpathy’s ā€œZero to Heroā€ series as a roadmap.

  • Reproduce the nanochat repo from scratch to understand how systems like ChatGPT actually work.

  • Learn in public, but without the pressure of perfection. I’ll share my journey here on this blog, explaining concepts like a beginner, because in many ways, I am one again.

Long-Term Focus: African Language AI

One of my long-term goals is to apply what I learn to speech recognition (ASR) and text-to-speech (TTS) for African languages, especially Fongbe. Building high-quality datasets and models for local languages isn’t just a technical challenge, it’s a cultural mission. I hope that by rebuilding my skills, I can contribute more meaningfully to that effort.

Closing Thoughts

This blog marks a new chapter, a return to curiosity, code, and consistency. If you’re someone who also stopped coding for a while, maybe you’ll see a bit of yourself in this journey. Let’s rebuild together.

Stay tuned. It’s good to be back. šŸ§ šŸ’»