30 October 2025 12:15 - 12:45
Interactive panel | Human oversight in LLMs: Building for safety, alignment, and trust
Aligning large language models to specific objectives demands fine-tuning pipelines that are both precise and scalable.
In this session, panelists will unpack the technical architecture behind these pipelines - covering data selection, model adjustment techniques, and ongoing alignment monitoring - while tackling challenges like preserving generalization and minimizing bias or drift.
30 October 2025 12:15 - 13:00
Interactive panel | Explainability & transparency in autonomous agents
Trustworthy autonomy starts with transparency - but making complex agents interpretable is a hard technical problem.
In this session, panelists will dive into the latest methods for surfacing how agents reason and act, tackling challenges like bias detection, traceability, and balancing explainability with performance. Learn how teams are building systems that stay accountable without slowing down.