Join Aishwarya Srinivasan and Arvind Narayanamurthy for a hands-on workshop on Fine-Tuning LLMs, how it works, when it’s worth doing, and how to run an end-to-end fine-tuning workflow without getting lost in jargon.

Fine-tuning can be high-leverage, but only in the right situations. In this session, we’ll dive in-depth into LLM fine tuning, build intuition for when to fine-tune, which approach to use, and how to tell if it genuinely improved performance.

This is a beginner to intermediate level session, designed for anyone working with LLMs who wants a grounded understanding of fine-tuning. Basic familiarity with LLM APIs and Python is recommended

Note: This workshop session will be recorded, and all participants will receive the recording post-event

What You’ll Learn:

The "Why" and "When" of Fine-Tuning: We’ll demystify the hierarchy of model training, from pre-training to post-training. You’ll learn to identify the "sweet spot" where fine-tuning pays off and, more importantly, when it’s actually the wrong tool for the job.

The Methods: From SFT to RLHF: We’ll break down Supervised Fine-Tuning (SFT) and provide a clear primer on Reinforcement Learning. You’ll understand how DPO (Direct Preference Optimization) and RLHF fit into the pipeline and when to use an "LLM-as-a-Judge" to scale your feedback loops.

Live Lab: End-to-End Fine-Tuning Demo: A practical, "no-fluff" walkthrough of a fine-tuning workflow. We’ll go from raw data to a specialized model, showing you the code and the logic behind each step.

Overcoming Production Bottlenecks: Training a model is one thing; making it work in production is another. We’ll discuss task-fit metrics, the "data quality levers" that actually move the needle, and the common pitfalls that cause fine-tuned models to fail in the wild.

Open Q&A & Insight Sharing: We’ll wrap up with an interactive session to troubleshoot your specific use cases, discuss the current tooling landscape, and share insights on where the field is heading.

By the end of this 90-minute session, you won’t just have a theoretical understanding of fine-tuning, you’ll have a clear, actionable framework for building specialized models that outperform "out-of-the-box" solutions.

The Gen Academy

LLM Fine-Tuning