Vai al contenuto
AI.info

events

Evaluate Coding Agents with Confidence: Tests, Trajectories & SambaEval

Un webinar gratuito in diretta sulla valutazione degli agenti di programmazione con pass@k, la valutazione delle traiettorie, il tracing e la selezione dei modelli SambaEval.

Evaluate Coding Agents with Confidence: Tests, Trajectories & SambaEval | Eventi sull’IA

In inglese

Starts
2026-11-18 19:00 UTC
Organizer
Data Science Dojo – Seattle
Location
Online event

Part four of Data Science Dojo’s SambaNova coding agent series is a free live webinar led by Kwasi Ankomah, Lead AI Architect at SambaNova Systems. It walks through an evaluation harness for coding agents that uses test outcomes as ground truth: pass@k across a task set, trajectory evaluation, and tracing for inspecting runs. It also covers SambaEval for choosing a SambaNova model for the executor role. The session is aimed at developers and researchers building or assessing coding agents.

Event website