ARIA, how we built autoresearch with autoresearch
Zubin Aysola
- When
- Thursday, July 211:10 AM – 11:30 AM · 20 min
- Where
- Expo Stage 2 NWSan Francisco, CA · imported from ai.engineer's public schedule feed
About this session
ARIA is an end-to-end auto research and AI research product that improves models, launches training jobs, and agents alike. We used ARIA along with a sophisticated evaluation framework we're calling the WBAF, Weights and Biases Agent Factory, to build itself. ARIA reads its own production traces, improves its own prompts, tools, skills, and other effects to solve customer challenges. In this talk, we dive into the evaluation framework, how we built a sophisticated reinforcement learning style environment over the Weights & Biases product, and how we scaled from zero to one to a full team working in parallel on improving an agent.
Speaker
Senior Software Engineer Weave, Weights & Biases by CoreWeave
For developers: this programme is open data — JSON, iCal, schedule XML and an MCP endpoint.Show endpointsHide
- JSONEvery published session and speaker, in one request./aie-worldsfair-2026-import/feed.json
- iCalSubscribe in Google, Apple or Outlook Calendar./aie-worldsfair-2026-import/feed.ics
- Schedule XMLfrab / pentabarf — the format conference apps import./aie-worldsfair-2026-import/feed.xml
- MCP + RESTPoint Claude at the programme. OpenAPI 3.1 included./agents
No key, no signup, CORS open. Everything here is generated from the same data the organisers edit.