Welcome!

Articles , Thoughts, Study Surveys

• Sapolsky Lecture Series:
  • Sapolsky Lectures: Lecture 1
• Automation Risk, Measurement & the Shape of Work

• Training_LLMs
  • Model Compression
  • Deepseek-LLM
• Basic RL for LLMs
  • GRPO - WIP!
    • GRPO Code slides
    • GRPO Code blog
    • Modal setup for GRPO
• DSPy
• [LEAN Provers]
  • Lean Conjecturer
    
  • a LEAN RL REPO
• Mojo GPU Puzzles
• NBDistributed

Papers , Mentorships & Collaborations

Authorships

• Thinking Before Constraining: A Unified Decoding Framework for Large Language Models
• IEEE SaTML 2026 : Membership Inference Attacks for Retrieval Based In-Context Learning for Document Question Answering
• Neurips 2025 LockLLM: Differentially Private In-Context Learning with Nearest Neighbor Search
• Certainty-Guided Reasoning in Large Language Models: A Dynamic Thinking Budget Approach
• Chip Thickness and Microhardness Prediction Models during Turning of Medium Carbon Steel

Acknowledgements

• Neurips 2024: Practical Differentially Private Hyperparameter Tuning with Subsampling
• (White Paper) Advancing AI: Networks that Self-Operate
• Location Theory & Decision Analysis

Mentorships

• Mitigating Catastrophic Forgetting in Task-Incremental Learning for Large Language Models
• Using log sequence representations and transfer learning in classification and novelty detection
• evaluating performance of generative models for time series synthesis
• Generative Models for Irregular Sequential Data