AI Engineer Sydney is the Sydney edition of the AI Engineer conference, run with Web Directions. Over two days at the Hilton Sydney, more than 800 AI engineers, software engineers, and technical leaders cover what it takes to ship AI systems in production: agent pipelines, evaluation frameworks, tooling, and case studies backed by production metrics. The program runs three tracks, AI Engineering, SWE and agentic coding, and AI Leadership, with a hands-on workshop day on December 9.
Theodoros Galanos, Forward Deployed Engineer at Nomic, will present The Benchmark Ends. The World Doesn't: Building Persistent Engineering Environments for Continual Learning. Before joining Nomic, Theo was Generative AI Group Leader at Aurecon and co-authored AEC-Bench, Nomic's open benchmark for AI agents on built-environment tasks. Track and time will be added here once the conference publishes the schedule.
The talk
A benchmark task ends when the answer comes back, but a live project keeps going. Revisions, RFIs, and handovers carry earlier decisions, and a missed comment cannot be patched after the pour. Theo will cover environments that hold that history, so the next agent inherits the decisions and open follow-ups already on the job.
Meet with us in Sydney
If you are attending, book time with the Nomic team to talk through:
- How agents review drawings, submittals, and specs against project documents, with every finding pointing back to its source
- What it takes to put domain-specific AI to work inside an engineering firm, from first files to a live project
- How we benchmark agents on drawing navigation, spec-drawing sync, and submittal review with AEC-Bench
