Future Skills For Nepal: Cracking Open A Large Language Model

Join SAGEA and HamroCSIT for a hands-on session on how large language models are built, trained, and refined for real-world use. Designed for students and builders, this session breaks down the core ideas behind modern AI systems without requiring prior experience.

We’ll cover how LLMs work under the hood, from architecture and training to how they learn from data at scale. The session then moves into advanced post-training techniques, including domain adaptation and instruction tuning, showing how raw models are shaped into practical systems.

We’ll wrap with a short Q&A to explore how you can start building with these concepts.