Project Background

Databricks wants to unveil a customer-facing technical session and live demo at Data + AI Summit for a new Mosaic AI agentic analytics workflow built on Databricks Agent Framework, Databricks Vector Search, Foundation Model APIs, Model Serving, Unity Catalog, Delta Lake, MLflow, and DBRX. The content must work for a mixed audience of AI engineers, data platform leaders, and executives, while still being technically credible enough to support follow-up field conversations and customer adoption.

You are the AI Engineer responsible for driving execution across a 9-person virtual team: 3 AI engineers, 2 solutions architects, 1 product marketer, 1 designer, 1 developer advocate, and 1 legal/compliance reviewer. The summit is 8 weeks away, and the session abstract is already accepted. Leadership expects the talk to generate reusable launch assets, not just a one-time presentation.

Key Stakeholders

The VP of Product wants a polished narrative that highlights Mosaic AI and Databricks Foundation Model APIs. The field CTO wants a technically deep demo with real Spark and PySpark pipelines, MLflow Agent Evaluation, and measurable RAG faithfulness and groundedness results. Legal requires all customer examples to use approved synthetic or anonymized Delta Lake datasets under Databricks Unity Catalog. Product marketing wants a simple story that can be repurposed into a blog, hands-on lab, and sales enablement deck.

Constraints

• 8 weeks to conference day
• $95,000 total event-content budget
• No additional headcount
• Demo must run live in under 7 minutes with a 2-minute fallback path
• Only one shared solutions architect is available 30% of the time
• Final abstract and title can only be revised once, by the end of Week 2

Complications

1. The latest internal agent demo uses an unapproved customer support dataset and cannot be shown externally.
2. The Mosaic Vector Search index refresh pipeline is unstable above 40M documents, creating risk for a live demo.
3. A PM wants to add a fine-tuned mosaicml model story in Week 5, which would likely displace evaluation and rehearsal time.

Your Task

1. Build an 8-week execution plan with milestones, owners, and dependencies.
2. Define how you would balance technical depth vs accessibility for a broad conference audience.
3. Propose launch assets, review gates, and rehearsal criteria for the session and demo.
4. Identify the top execution risks, escalation points, and fallback plans.
5. Specify success metrics for the talk, demo, and post-event reuse.