Free Lesson
What Makes a Good Search Agent?
Part of AI Product Engineering
60 min
Jul 13, 2026 2:00 PM
Virtual (Zoom)
In this video
What you'll learn
Build benchmarks that test search agents
How to construct benchmarks that hold up, using BrowseComp-Plus and MAST to evaluate agents on hard, agentic browsing tasks.
Generate training data on a budget
How ORBIT synthesizes multi-constraint queries cheaply, so you can train search agents that generalize beyond one narrow domain.
Read agent trajectories to find what's wrong
Why a search agent's trajectory explains its performance, with a demo of the Hawkeye tool for analyzing them.
Why this topic matters
Search agents query, retrieve, and reason over external sources to answer knowledge-heavy questions. Nandan will show you how to build benchmarks that measure search-agent quality, how to synthesize training data on a budget, and how to use trajectory analysis to debug these agents.
You'll learn from
Nandan Thakur
Creator of the BEIR and MIRACL benchmarks; PhD, University of Waterloo
Hamel Husain
ML Engineer with 20+ years of experience