Company / People

The people building Litil.

Litil brings legal judgment, model training, product, and production software into the same lab.

We are building the layer that should remain useful when a model changes: the work, the agent, the environment, the evaluation, and the record of what happened.

Person / 01Law / models / environments

Aaron Kelly

Aaron Kelly is a technology attorney and AI builder at LiTiL Labs, where he develops open models, legal agents, reinforcement learning environments, and post-training systems for legal work. His approach draws on experience in commercial contracts, corporate governance, and legal leadership at Edge & Node and Zellic. He focuses on turning legal documents, decisions, and practitioner feedback into better training data, evaluations, and AI systems that organizations can own and improve.

Person / 02Technology / systems / product

Garnett Gilchrest

Garnett Gilchrest is a technology executive with experience leading software teams, product development, and enterprise architecture. His career includes serving as CTO at Solvace and Executive Vice President of Information Technology at Multiview. At LiTiL Labs, he brings that experience to the software and infrastructure surrounding legal AI models, helping translate the lab’s work on agents, environments, and models into practical products for legal teams.

Person / 03Law / software / AI systems

Sol Irvine

Sol Irvine is a transactional lawyer and software developer with more than twenty-five years of experience in complex technology transactions. His career includes Shaw Pittman, Morgan Lewis, and co-founding Yuson & Irvine. He also founded Concertiv, serving as General Counsel and CTO. Today, he builds AI tools for transactional lawyers, including Augustus for contract analysis and Wargame for simulated negotiations. At LiTiL Labs, he brings legal practice and software development experience to systems that reflect how deals are actually negotiated.

Person / 04AI / quantitative research / commercial strategy

TJ Barton

TJ Barton is the co-founder and CEO of DXRG, where he focuses on building frontier AI models for trading applications. Since 2023, he has brought his background in quantitative research and commercial strategy to applied AI development. Previously, he served as Vice President of Category Revenue Growth Management at Keurig Dr Pepper, leading commercial strategy for an $8B+ consumer-products portfolio, with experience spanning pricing, trade strategy, and commercial analytics. TJ holds a B.S. in Physics from MIT, and conducted computational physics research at MIT’s Plasma Science and Fusion Center. He brings this combination of scientific rigor, large-scale business decision-making, and model-training experience to his work supporting legal AI.

Why Litil

The learning should stay with the work.

Legal teams give AI systems documents, instructions, corrections, decisions, and outcomes. Litil is building a way to keep that record, understand what happened, and improve the right part of the system without starting over when the provider changes.

The lab is open about the work we can show. Models, code, evaluations, and training reports belong in the public record when they are ready. Work that is private stays described at the level the evidence supports.