Teams

Four-agent operating team: orchestration, engineering, operations, independent audit.

Bilingual support pod: frontline resolution plus localization.

Open-source AI software developer with sandboxed runtime — ICLR 2025.

Non-agentic 3-stage pipeline — 40.7% Lite / 50.8% Verified with Claude 3.5 Sonnet.

Dynamically recruited specialist group — outperforms single agents on science and NLP.

Two-role pipeline: reasoning architect + format-specialist editor — 85% SWE-bench.

Internal fleet-scale Java 8/11→17 upgrades — >4,500 dev-years saved, $260M.

Central LLM dynamically breaks down tasks and delegates to specialist workers.

Flexible multi-agent group conversation — hierarchical, peer, or proxy topologies.

Execution + Guard agent pair — GAIA score timeline, latest 67.89 Pass@1.

Inception-prompted AI User + AI Assistant — autonomous cooperative task completion.

5-role sequential pipeline — 22,949 tokens, 148s per software task.