I’m a researcher and educator exploring how modern AI systems think, attend, and align.

My current work builds a bridge between algorithmic research and practical systems engineering, through a series of paid, research-grade courses designed for advanced machine learning engineers:

• Which Attention is All You Need?
  Deep exploration of modern Attention architectures — Linear, Sparse, and Hybrid — their scaling laws, memory tradeoffs, and why even industry systems like MiniMax and DeepSeek still return to Full Attention under certain constraints.
• RLHF, GRPO, and Post-training Optimization
  A detailed walkthrough of how alignment training actually works — from human feedback modeling to gradient-based preference optimization — with hands-on system design and reproducible code.
• From Algorithms to Systems
  How commercial LLMs balance compute, architecture, and alignment. A view into real-world deployment tradeoffs and the future of efficient inference.

These are deep, technical, and practitioner-oriented programs — crafted for senior researchers and engineers who want to go beyond framework-level abstraction and understand how to truly design models.

Sponsorships directly fund the continued research, experiments, and materials behind these paid programs.

🌐 Learn more or join upcoming sessions at my Youtube channel.