Announcing our Series C with $110M in total funding. Read more →.

AI After Hours London: AI in the Physical World

February 24, 2026 | Encord x Anyscale x Dawn Capital

AI After Hours London: AI in the Physical World

Speakers

Oana Andreea Jinga

Oana Andreea Jinga

Co-Founder @ Dexory

Oscar Mendez

Oscar Mendez

Head of AI & Data Science @ Locus Robotics

Peter Overbury

Peter Overbury

Senior Computer Vision Engineer @ VisionTrack

Rui Campos

Rui Campos

VP of Product @ Automata

About this Event

AI After Hours brought together more than 100 attendees at Dawn Capital HQ for an evening focused on the future of Physical AI. The panel discussion explored the challenges of moving AI from controlled environments to real-world deployment, the role of vision and sensing in autonomous systems, and what it takes to build safer, more reliable Physical AI systems. Moderated by Eric Landau (CEO, Encord), the conversation covered everything from world models and multimodal data to the path toward general-purpose Physical AI.

Stay updated on upcoming events

Join our mailing list to hear about future AI After Hours events, panels, and exclusive content from industry leaders.

Event Highlights

AI After Hours was back in London on February 24, 2026! Hosted at Dawn Capital HQ in collaboration with Anyscale, more than 100 attendees packed the room for an evening of networking and insights as well.

The Panel: AI in the Physical World

Our expert panel featured Oana Andreea Jinga (Dexory), Oscar Mendez (Locus Robotics), Peter Overbury (VisionTrack), and Rui Campos (Automata), moderated by Eric Landau (Encord). The conversation covered:

  • The Brain and the Senses: How world models are becoming essential for Physical AI, and the ongoing debate around vision-only systems versus multimodal sensing (LiDAR, Radar, tactile sensors).
  • From Lab to Real World: The challenges of moving from controlled demos to customer-ready deployments, and why embodied data is so messy to work with in production.
  • Building Safer Systems: What it means to define "safe" for Physical AI, and how teams handle edge cases and physical phenomena that are harder to model than expected.
  • The Path to General Purpose AI: Whether a single AI model can work across cars, drones, and warehouses, or if winners will always build full-stack solutions for specific niches.

Upcoming Events

Join us for our next in-person events.