About the Event

🚀 Join us for an exciting workshop focused on enhancing retrieval agents! Many retrieval agents operate quietly but often produce incomplete or incorrect results. In this hands-on session, you'll learn how to construct and assess an intelligent agent that can recognize when its retrieval step is suboptimal.

🛠️ Workshop Highlights:

• Start with a prebuilt, Qdrant-powered agent
• Conduct evaluations, inspect traces, and pinpoint failure modes
• Implement in-loop checks to enable the agent to decide on rewriting queries, searching again, or stopping
• Learn to analyze practical retrieval-quality signals such as:
  • Low confidence in top results
  • Tightly clustered scores
  • Weak rank gaps
  • Discrepancy between keyword and vector retrieval
• Improve the loop using these signals and measure actual improvements in answer quality, retrieval efficiency, cost, and time-to-success.

🎒 No Pre-work Required! Every participant will receive a ready-to-use VM equipped with the environment, data, agent code, Qdrant setup, traces, and evaluation tools.

🗓️ Agenda

• 5:00 – 5:40 PM: Arrival, Registration & Networking
• 5:40 – 6:00 PM: Introduction
• 6:00 – 6:20 PM: Nikhil Pareek, Founder & CEO, Future AGI - Closing the Loop: Engineering Self-Improving AI Agents
• 6:20 – 7:20 PM: Dylan Couzon, DevRel Engineer, Qdrant - Building Self-Correcting Agentic Retrieval Loops
• 7:20 – 7:50 PM: Amanda Hua, Founder, azenticbot.ai - When Retrieval Fails: Governing Agent Decisions Under Uncertainty
• 7:50 – 8:30 PM: Q&A
• 8:30 – 9:00 PM: Wrap-Up & Networking

What You’ll Learn

• Evaluate retrieval agents beyond final answer correctness
• Inspect traces to identify retrieval-driven failures
• Use live retrieval signals for better decision-making
• Detect weak retrieval with score confidence, variance, and rank gaps
• Understand how query rewriting can impact retrieval
• Measure improvements step-by-step
• Balance answer quality, latency, token costs, and tool-call expenses
• How Qdrant enhances retrieval for adaptive agent systems

Who Should Attend

This workshop is ideal for AI engineers and agent developers looking to transcend the basics of "retrieve and generate," seeking effective techniques to evaluate and enhance agent functionality. This is best suited for those who have experience with foundational agent workflows and desire improved debugging tools for retrieval failures, query rewriting, and decision-making.

🔗 Reserve Your Spot Here!

Location: Digital Jungle SF
Seats are limited!