london
I like coding, optimising and learning. My main focus is on full-stack development with Next.js, but I'm also interested in AI and neural networks. In my free time I play tennis and hit the gym.
Products I built, shipped and maintain for real teams.
Full-stack ERP for companies; manages projects, invoicing, scheduling, more features listed in blog post. Scaled to 2000+ projects and 20+ daily users, saving hundreds of hours on manual tasks.
2 companies · 2,000+ projects · 20+ daily users
Multi-tenant secure file sharing SaaS with pricing tiers that supports chunked uploads, fast downloads and has audit logs.
5+ companies · 6,000+ users · 500 GB+ stored
Guided builds where I worked through unfamiliar ideas in code.
Worked through character-level language modelling, comparing a count-based bigram model with a gradient-based neural approach.
bigrams · probability · gradient descent
Worked through a small autograd engine and neural network library to understand computation graphs and backpropagation.
autograd · backpropagation · computation graphs
Followed a raycasting build to understand how 2D geometry, rays and projection can create a simple 3D scene.
ray casting · trigonometry · canvas
Worked through a NumPy neural network trained on MNIST and explored how it could power an interactive web demo.
neural networks · backpropagation · MNIST