Betting on Open Source: Post-Training Models in Production with EliseAI & Baseten

Join EliseAI and Baseten for an evening exploring how teams can adapt open-source models for specific use cases, deploy them reliably in production, and keep pace with a rapidly evolving ecosystem.

Engineers from both teams will compare strategies for adapting and improving open-source models: when fine-tuning and other post-training techniques are worthwhile, when harness engineering is the better approach, how to build accurate benchmarks and reliable ground-truth datasets, and how to hill-climb on evals without optimizing for the wrong signals. They’ll also discuss the performance, infrastructure, and reliability challenges that emerge at scale.

We’ll kick things off with a short technical discussion, followed by an open Q&A and networking with engineers, researchers, and builders from across the AI ecosystem. Come for the technical deep dive, stay for drinks and conversation with others building at the frontier of AI.

What to expect:

6:00 PM: Doors open, Pizza & Drinks 🍕🍺

6:45 PM: Panel: Post-Training Models in Production

7:15 PM: Open Q&A & Technical Deep Dive

7:45 PM: Networking & Community

9:00 PM: Doors Close

About us:

EliseAI builds AI automation for the housing and healthcare industries, which together make up 40% of American household spending. Our AI powers operations across the full customer and resident lifecycle—from initial inquiry and leasing through ongoing resident engagement, renewals, and other critical workflows. Today, one in six leasing apartments in the U.S. runs on EliseAI.

Baseten is the training and inference platform behind some of the most demanding AI workloads in production for companies like Abridge, Cursor, Notion, and OpenEvidence.