Hands-on workshop: Advanced model alignment for Healthcare

Synthetic data generation with NVIDIA NeMo DataDesigner, post-training and alignment with NVIDIA NeMo RL

How do you fine-tune a model when real-world medical data is scarce or sensitive? This workshop demonstrates how to bridge the gap. We will walkthrough how to generate synthetic data using NVIDIA NeMo Data Designer, then progressing through Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) using NVIDIA NeMo RL to tailor models for healthcare applications.

You will leave with practical techniques and insights to build and improve models for clinical tasks.

What to expect:

• Synthetic data generation: Use NeMo DataDesigner to build high-quality training data for data-poor medical niches

• Domain adaptation: Transform a general-purpose LLM into a purpose-built healthcare agent

• RL alignment: Fine-tune models to conform to specific healthcare protocols using NVIDIA NeMo RL

• Guided Implementation: Build and experiment in a structured environment with NVIDIA Solution Architects

Places are extremely limited, sign up now to avoid disappointment.

If you can’t join the workshop and/or would like to hear more, join the Nebius.Build/LON that afternoon.