Aller au contenu
AI.info

events

Evaluate Coding Agents with Confidence: Tests, Trajectories & SambaEval

A free live webinar from Data Science Dojo Singapore, part four of a series on SambaNova coding agents. Kwasi Ankomah, Lead AI Architect at SambaNova Systems, explains how tests provide ground truth…

Evaluate Coding Agents with Confidence: Tests, Trajectories & SambaEval | Événements de l’IA

En anglais

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

A free live webinar from Data Science Dojo Singapore, part four of a series on SambaNova coding agents. Kwasi Ankomah, Lead AI Architect at SambaNova Systems, explains how tests provide ground truth for coding agents and how to build an evaluation harness around that signal. The session covers pass@k across a task set, trajectory evaluation, full tracing so runs can be inspected, and using SambaEval to choose a SambaNova model for the executor role.

Event website