Session Description

In this in-person, half-day training, we’ll start with the motivation, organizational context, and management perspective behind (re)platforming search, then move into concrete implementation topics — evaluation, agentic search design, business rules, and content enrichment for AI consumption.

Part 1 — Motivation, Context, and the Management Perspective

• Why (re)platform search now?
• Starting point: assessing your organization’s search maturity
• Data readiness as a key prerequisite for building modern search platforms, and what differs between a search application team and a platform team in terms of data ownership and transformations
• Team structures: applying Team Topologies to search platforms
• Platform for Search: the main components and specifics of a modern search platform, and the principles that make it accessible and suitable for AI

Part 2 — Towards an Implementation

• Evaluation: where it fits — “offline” vs. as part of retrieval — with a quick recap of LLM-as-Judge and other evaluation approaches
• Designing for agentic search: interfaces, requirements on retrieval, one Search API vs. exposing individual components, and the query understanding component
• The role of business rules and how to manage them
• Auto-research
• The data side: content enrichment and making content understandable to agents

This part includes hands-on exercises applying these ideas to real design decisions.

Who should attend this training?

• Platform and architecture leaders
• Engineering managers
• Product leaders working with AI and data-driven systems

This session is held in person as a half-day format, alongside our companion training, “LLMs as Judges for Search Result Quality”, on the same day.