Optimize ARC-AGI-3 Kaggle Solution

I’m looking for an experienced ML engineer to optimize an ARC-AGI-3 Kaggle solution and achieve a top-5 position on the public leaderboard within 7 days of the agreed start date.

Budget: $800 USD, success-based. Payment will be made only if the agreed top-5 target is achieved by the deadline and all required deliverables are provided. No payment is due if the target is not achieved, unless otherwise agreed in writing before work begins.

Starting points:
https://www.kaggle.com/code/dfranzen/arc-agi-3-milestone-2-solution
https://www.kaggle.com/code/yukailiu/arc-agi-3-champion-solution

Deliverables:
- A working, reproducible Kaggle notebook.
- All modified source code and configuration files.
- A valid competition submission confirming the result.
- A brief explanation of improvements and reproduction steps.

Preferred skills: Python, LLM agents, Kaggle, SGLang/vLLM, quantization, and inference optimization.

Please include your Kaggle/GitHub links, relevant experience, proposed approach, and GPU requirements. Computing costs must be agreed upon before work begins.

The exact deadline, ranking verification time, and submission account will be agreed upon before starting. All work must comply with competition rules and open-source licenses.