Summary: How do data, business semantics, and context work together? How do you build a lightweight application? Where does an enterprise rollout begin? The speakers and agenda for the October 20 MIP launch are now live.
A product introduction is only the first step in getting to know an enterprise AI data platform.
How it organizes data and business semantics, which scenarios it supports, what it looks like in use, and how it gets rolled out inside an enterprise all need to be seen together, through architecture, demos, and real-world practice.
On October 20, MirrorShip will host the MIP launch event. Five speakers will cover product direction and technical design, walk through a hands-on demo of a lightweight application, and share methods and lessons from multimodal data lake retrieval and enterprise implementation.
Here's a look at what to expect.
What should data infrastructure deliver in the AI era?
As data starts serving AI, which existing capabilities still matter, and which needs have to be rethought? That question is the starting point for MIP's product direction.
Vincent Sun, CEO of MirrorShip, will share his thinking on the path from StarRocks to MIP, the data infrastructure needs of the AI era, and where MIP is headed. His talk explains why MirrorShip built MIP and the product decisions behind it.
How do data, business semantics, and context work together?
A business question can rarely be answered by querying data alone. What a metric means and what background a task requires also need to be part of the analysis.
Understanding MIP means looking past individual features to see how data, business semantics, and context fit together.
Andy Chen, CTO of MirrorShip, will lay out these relationships from a technical perspective and introduce the technical philosophy and architectural principles behind MIP. The product demo that follows builds on this design thinking.
Inside the product: how the architecture works
After product direction and technical philosophy, the next question is more concrete: what do these designs look like in the product? How do the three layers divide the work, which scenarios do they fit, and what applications can they support?
Sophia Dong, Head of Product for MIP, will move from architecture to hands-on use. She will cover the product definition, the three-layer architecture, and target scenarios, then show a lightweight application in action through a live demo.
This session looks at how MIP's capabilities are organized and applied to specific tasks, beyond a list of what the product includes.
From business tables to multimodal data lakes
The launch will also look beyond MIP itself to retrieval practices in the StarRocks ecosystem.
In large model training and autonomous driving, the goal goes beyond getting a query result. Teams need to filter training data and find the Qhard cases worth a closer look, which is where vector search and full-text search have to work together.
Zhou Kang, Head of the Open-Source OLAP Engine Team at Alibaba Cloud and StarRocks TSC Member, will share practices for multimodal data lake retrieval with StarRocks and Paimon. If you work on data filtering and hybrid search for AI, this talk is worth your time.
From understanding the product to starting implementation
Finally, the focus returns to the enterprise: what does it take to get data ready to serve agents?
Data onboarding, semantic modeling, and AI application delivery are connected pieces of work. Once teams understand the product, they also need a clear view of how that work gets done.
Li Bao, Big Data Architect at Guangzhou Haijie Technology, will share methods and practices for implementing MIP from an enterprise perspective. For teams planning similar projects, this talk shows how an application is built step by step, alongside what it delivers in the end.
Agenda
.png)
Join us on October 20
If you lead enterprise data or AI platform work, are exploring agents for business use, or follow the StarRocks ecosystem, register to watch.
From product direction to technical architecture, from live demo to ecosystem practice, the launch will give you a closer look at MIP and at how AI-ready data platforms are put to work.
- Date: Tuesday, October 20, 2026
- Time: 14:00–15:50 (GMT+8)
- Format: Online
Register on LinkedIn: https://www.linkedin.com/events/7515230959584305152/?viewAsMember=true