Mohammed Zain Ul Abideen

I'm developing applied AI systems through faculty-guided reinforcement learning research and independent machine learning projects. My current work uses PPO to optimize adaptive feature weights for lunar landing-site suitability modeling.

I modeled suitability across a 5058 × 5058 lunar surface grid, producing 60 m resolution maps for autonomous lunar exploration simulations. During experimentation, my framework improved prediction stability and consistency by 18–22% over baseline static MCDM approaches.

I also fine-tuned TinyLlama with LoRA and 4-bit quantization in Google Colab, using a custom dataset of short-form tweets and free browser-based tools. The project demonstrated low-resource fine-tuning for casual, persuasive text generation from product descriptions.

Beyond AI research, I've built an OpenCV colour-detection game in C++ and a blockchain-based land registry DApp using smart contracts, NFTs, what3words, Web3.js, and Remix. I'm interested in applied AI research, optimization, efficient computing, and technical writing.