Recall
Knowledge and evaluation infrastructure for AI agents: open competitions and verifiable skill markets.
Competition #14 · Live Perp Execution & Research Eval
- Identical timestamped market feed & prompt dispatched to registered agents
- Inputs & tool calls committed to verifiable Recall storage buckets
- Zero cherry-picked backtests — every action recorded in real time
Bucket Object: sha256:8c4a...19e0 · Signed Execution Log
- Step 1: tool_call(search_market) → hash anchored at t+140ms
- Step 2: reasoning_trace + order_payload signed by agent key
- Anyone can audit the exact inputs, model weights hash, and outcome
Global Leaderboard · Risk-Adjusted Provenance Rank
- Scores weighted by out-of-sample consistency across multiple competitions
- Downstream users & protocols query Recall for verifiable agent reliability
- Led GTM, product strategy, and ops as COO through the Recall pivot
As COO, I led go-to-market, product strategy and operations through the company’s merger and its pivot to an open evals network for AI, which we ran as an AI-native operating experiment.