Skip to content
InfratoApp

NVIDIA · Course

Adding New Knowledge to LLMs (self-paced)

Adding New Knowledge to LLMs is a full-day, hands-on, self-paced NVIDIA workshop on turning an open-source language model into a domain specialist, with an evaluation, synthetic data, LoRA fine-tuning and preference optimisation on a triage case study.

Open at NVIDIA ↗Common questions

Who it suits

  • ML engineers who need a model that follows local taxonomies and returns reliable schemas.
  • Teams deciding whether prompting, fine-tuning or preference optimisation is the right next move.

What it covers

  • Designing an evaluation that captures task behaviour, measurable success and failure modes before training.
  • Using evaluation evidence to choose between prompting, synthetic data, supervised fine-tuning and preference optimisation.
  • Generating targeted synthetic data with a teacher model and NVIDIA NeMo Data Designer.
  • Training a LoRA adapter with supervised fine-tuning, then improving responses with Direct Preference Optimization.

Common questions about Adding New Knowledge to LLMs (self-paced)

How is it taught?
As an end-to-end pipeline on a fictional humanitarian-logistics case study, with a capstone in a customer-support domain.
Does it cost anything?
NVIDIA sells the self-paced course with Buy Now and Redeem Code options, and says online purchase is supported in select locations. Its page states the price, which changes, so check it.
Which model does it use?
NVIDIA describes measuring a stock 8B open-weight model and then improving it.

More like this

Keep learning on Infrato

Where this comes from

Written from NVIDIA’s own pages, read on 2026-10-14. Content, prices and availability change, so the provider’s page is the source. We are not affiliated with NVIDIA and this is not an endorsement.