2 posts
Optimizing Agents
Why we built AgentGrad.
The library behind our agent-optimization engagements, out of stealth: a PyTorch-shaped training loop for agent runtimes, with textual gradients, batch-level updates, and constraint-preserving steps. Why we built it instead of adopting what existed.
agents·11 min·Aug 2026
From the Field
Legal tool calls, decomposed.
A field note on legal tool-call optimization: thin tool routing, per-tool argument builders, and failure-driven prompt/program optimization for higher argument accuracy with less context pressure.
legal·11 min·Jul 2026coming soon