Registration closed
This workshop is at capacity and registration is now closed. Thank you for the strong interest. If you would like us to follow up afterward with materials and details on the next session, leave your details below.
Request a follow-upAbout the workshop
In May we showed one autonomous loop designing one component. That work is now published (arXiv:2606.00915) with an open framework. Since then we have taken agents through the rest of the flow: system-level simulation, placement, routing, and verification, ending in a manufacturable layout.
This workshop covers both halves. Session 01 is the device, Session 02 is the chip, and in Session 03 you run it yourself. We also cover where human intervention is needed, so you understand the limits of agentic chip design.
Session 01
Tackling problems across active and passive component design with Tidy3D. We run the loop on splitters, grating couplers, PN-junction phase shifters, and RF electrodes, and show what the agent needs to be given before it can do any of it.
Session 02
System-level simulation, placement and routing, and verification, ending in a manufacturable layout. One written specification becomes a numeric target for every component, then a routed, DRC-clean chip.
Break
Lunch is provided. A good moment to get your laptop set up before the hands-on lab.
Session 03
You run the workflow. Point a single agent at one component, or give an orchestrator a specification and let it dispatch its own team, with Flexcompute engineers in the room. Bring your laptop.
In practice
Follow-up
The workshop is full and we are no longer accepting registrations. Leave your details and we will follow up afterward with materials, and let you know when we run this again.
Created by engineers from MIT and Stanford, Flexcompute is redefining simulation technology to help innovate faster, cut costs, and reduce risks. We're making hardware development as easy as software.
.gif)
