Posts on Saheb's AI Journey

On failing at my first Kaggle competition (3/n)

Sat, 27 Jun 2026 00:00:00 +0000

In the second part we talked about transformer heads and features.

In this part, we will dive into credit assignment and reward hacking problems that I faced.

Giving +1 reward when you win a game and -1 when you lose is clean and eventually you want to reach that state where it suffices but before you get there, you need more signals, i.e. dense rewards to teach the agent the behavior you want.

On failing at my first Kaggle competition (2/n)

Fri, 26 Jun 2026 00:00:00 +0000

In the first part, we discussed self play and the equilibrium our agent gets stuck in.

In this part, we will talk about features that model takes as input, and how model uses them to learn strategies.

Theme and tone will be the same, more on what not to do.

In Orbit Wars, features fell into 4 categories, planets, fleets, global, and pairwise features.

Planet features will tell the model everything about the planet, its position, owner, radius, production, orbit info, capture cost, distance to nearest owned planet, etc.

On failing at my first Kaggle competition (1/n)

Thu, 25 Jun 2026 00:00:00 +0000

I had a lot of fun building an AI agent for Orbit Wars.

But I failed, according to my own standards, and naive high expectations.

I wanted to be in top 10. I didn’t even end up in top 1000.

The final scoreboard: a rating of 995, ranked 1060 out of 4729. Top 25%, but nowhere near the top 10.

Getting there took over 90 training runs across a month, and about $700 worth of GPU compute across GCP, AWS, and Jarvis Labs.

The Day My AI Editor Went Silent: Debugging Git Worktrees, Phantom Configs, and Lost Weights

Mon, 30 Mar 2026 00:00:00 +0000

Building reinforcement learning environments is already a complex exercise in managing state, rewards, and training loops. You expect the agent you are training to act unpredictably. You don’t expect the AI agent writing your code to completely flatline.

While working on some recent RL projects, my AI code editor—Google Antigravity—just gave up. Every time I hit send, it would emit a single beep. No error message. No spinning loader. Just a beep and a dead UI.

Reinforcement Learning with Snake (1)

Sun, 01 Mar 2026 00:00:00 +0000

Reinforcement learning is interesting because it is intuitive and it aligns with how we learn naturally. You touch hot water, you feel the burn, you don’t touch it again. You learn by trial and error; you learn from experience. By interacting with the world, you get feedback in the form of reward or punishment, and that defines your behaviour.

You notice there is a knob next to the water tap. It has a red arrow and a blue arrow on it. You press it — nothing happens. You pull it — nothing happens. You turn it and voila, the water starts again. You dare to touch it again, and this time it’s slightly less hot … and that’s how you learn by exploring. Either you can do this yourself, or someone teaches you and robs you of the opportunity of figuring it out on your own.