Raaghav P
AI engineer, building the agent harness at 
Figr, a design agent for product teams.
Lately that means questions like:
- Context management: How do you give an agent the right design context for the task in front of it?
- Memory: What should an agent remember between sessions, and how do you keep that memory from going stale?
- Evals: How do you tell whether a design it produced is actually good?
If you're thinking about the same problems, I'd like to hear from you — find me on GitHub, X, LinkedIn, or email.
Blog
| Coming soon |
Projects
| Coming soon |