


Closing the Loop
An ongoing series for those pushing real intelligence into the physical world.
Closing the Loop is a speaker series and community hangout for folks working at the edge of robotics, control, embodied AI, and physical intelligence.
Robotics is hard: sensors lie, actuators drift, models fail. Embodied intelligence pushes every layer of the stack—hardware, perception, actuation, control, and failure recovery—under real-world constraints. This is a space for practitioners to learn from each other, swap ideas, and build kinship in a community that understands the grind.
Each session features a handful of short, sharp talks from people building systems that sense, act, and adapt in the physical world. The focus is on sharing real challenges, hard-earned progress, and lessons learned—not polished demos or marketing decks.
There will be pizza, soda, and time to mingle.
Small and intimate. No keynote energy. Just smart people working on tough problems, sharing what they know.
Hosted by Sauron, where we’re building the autonomy stack for the home.
Agenda
5:30 PM – Doors open
6:00 – 7:00 PM – Invited talks
7:00 – 8:00 PM – Mingling + conversation
Talks
Karthik Lakshmanan from Zipline
Scaling to the Sky: Lessons from 100 Million Miles of Autonomous Delivery
With over 100 million miles flown across multiple continents to complete 1.5 million commercial deliveries, Zipline builds and operates the world’s largest autonomous logistics system. Achieving this scale required not just robust robotics and autonomy, but also the right engineering and organizational investments to ensure real-world reliability.
This talk shares key lessons we've learned along the way: investing in data-driven techniques to handle the incredible diversity of the real world, leveraging simulation and software-in-the-loop testing to accelerate iteration, integrating early and often, incorporating customer feedback from day one, and building trust across a complex ecosystem of regulators and partners.
Szymon Jakubczak from Physical Intelligence
What would it take for a robot policy to perform complex tasks in previously unseen uncontrolled environments? This talk will describe how Physical Intelligence made a bet on real-world data diversity and what we learned in the process.
Brannon Smith from Sauron
About Sauron
Sauron is building spatial intelligence for the home—systems that identify, understand, and actively deter threats. Our mission is to empower homeowners with autonomy.
💡 Curious to learn more?
🔗 Follow Sauron on LinkedIn
🚀 We’re hiring—join our team.
