
ALT AI Engineer New York Classic session guide in black and champagne with the original event logo and official portraits of all five speakers. Your Tuesday in New York. October 13, 2026. Four stops for finance and agent evals. 9:00 AM ET, Main Stage. Give your agent the research. Aaron Linsky and Sam Green, Bridgewater. 10:40 AM ET, Simons Room. Learn from banking transactions. Freddy Lecue, Wells Fargo. 11:05 AM ET, Turing Room. Build the eval dataset. Nathanael Weill, Air Canada. 3:20 PM ET, Turing Room. Catch what your fix broke. Sofia Jakovcevic, Arize AI. All session times Eastern. Sheraton New York Times Square. Full agenda in the post.

ALT Source-native Figure 3 from HANDBOOK.md. A plotted comparison uses circles for strict pass@1 and diamonds for pass@1 (N−1), which allows one failed criterion. The diamond scores rise above the circle scores across the June 2026 configurations, including Opus 4.8 (max) from 21.9% to about 46%, showing how often a single policy miss separates a near-success from a strict pass.



























