Training models one step at a time
I'm a Research Scientist at Salesforce (joined via Tenyx acquisition). I currently work on the fundamentals of Large Language Models, focusing on training methodologies, understanding, and agentic applications.
What's New
- Dec 2025Blog post with Silvio on the Agent-to-Agent Semantic Layer
- Nov 2025Paper out on arXiv: Echoing: Identity Failures when LLM Agents Talk to Each Other
- Nov 2025Paper on arXiv: Convergence dynamics of Agent-to-Agent Interactions with Misaligned objectives
- Oct 2025Adam presented our work on A2A at Dreamforce 2025. Check out the youtube video.
- May 2025Fun project: AGI Is Coming... Right After AI Learns to Play Wordle
Selected Publications
View allS Shekkizhar, R Cosentino, A Earle, S Savarese · arXiv Preprints, 2025
As large language model (LLM) based agents interact autonomously with one another, a new class of failures emerges that ...
R Cosentino, S Shekkizhar, A Earle · arXiv Preprints, 2025
We develop and analyze a theoretical framework for agent-to-agent interactions in a simplified in-context linear regress...
S Shekkizhar, R Cosentino · arXiv Preprints, 2025
This paper investigates multimodal agents, in particular, OpenAI's Computer-User Agent (CUA), trained to control an...