Post-Training at the Frontier: How You.com, NVIDIA, and CoreWeave Improved Nemotron’s Agentic Search & Browsing Capabilities

In this session, You.com’s engineering team will share their work with Nvidia and CoreWeave to post-train Nvidia’s Nemotron 3.5 Lightning on CoreWeave using Dynamo RL rollouts in NeMo gym with You.com’s web search API. We’ll walk through the technical approach, which enabled the team to rapidly iterate and drive measurable gains on BrowseComp, a benchmark for agentic web search and browsing. We will cover topics ranging from dataset preparation, supervised fine-tuning, and reinforcement learning to CoreWeave’s infrastructure for post-training. We’ll end with a blueprint for post-training and improving LLMs on real-world search tasks as well as a look at what’s next.