Blog

Programming & Code

+2

RLHF Training at Scale with DeepSpeed-Chat

Feb 17, 2026

•

9 min read

RLHF Training at Scale with DeepSpeed-Chat

Efficient RLHF Training for Large Language Models with DeepSpeed-Chat, PPO, and LoRA

Dave Berry

Blog

+1

New Insights into the Inner Workings of In-Context Learning

Feb 15, 2026

•

8 min read

New Insights into the Inner Workings of In-Context Learning

A Meta-Optimization Perspective on How Transformers Adapt Without Updating Weights

Dave Berry

Programming & Code

+3

Bite-Sized Bayesian: A Mini-Tour of Bayesian Inference, from Coin Flips to Neural Networks

Feb 14, 2026

•

5 min read

Bite-Sized Bayesian: A Mini-Tour of Bayesian Inference, from Coin Flips to Neural Networks

How priors, posteriors, and probability distributions help us reason under uncertainty

Dave Berry

Blog

+1

Fine-Tuning Models? Think Surgical Precision, Not Sledgehammer

Feb 12, 2026

•

4 min read

Fine-Tuning Models? Think Surgical Precision, Not Sledgehammer

Why selective layer tuning can outperform full fine-tuning under distribution shift

Dave Berry

Blog

+1

Demystifying AI Foundation Models: A Comprehensive Guide to Large Language Models, ChatGPT, and Beyond

Feb 9, 2026

•

26 min read

Demystifying AI Foundation Models: A Comprehensive Guide to Large Language Models, ChatGPT, and Beyond

How large language models work, why they matter, and where they’re headed

Dave Berry

Dave Berry

Scientist, Musician, Entrepreneur — Exploring science, creativity, technology, and life through research, music, and thoughtful inquiry.

Disclaimer: The views expressed on this site are my own and do not represent those of any employer or organization I’m affiliated with. Content is for informational and educational purposes only and is not medical, legal, financial, investment, or professional advice. Always consult qualified professionals regarding your specific situation.