TileDB becomes Tile.ai as founder hands CEO job to George Llado

Why it matters

Papadopoulos is separating technical vision from commercial execution as Tile.ai tries to sell a cross-system governance layer into regulated enterprises.

Stavros Papadopoulos is handing Tile.ai's CEO job to George Llado and returning full-time to product, pairing a leadership change with a new pitch for the database technology he began building at Intel Labs and MIT.

TileDB announced the changes on September 16th, rebranding as Tile.ai and putting a product called the Enterprise AI Data Substrate into private preview. Existing investor Two Bear Capital led a strategic financing, with Amgen Ventures participating. Tile.ai announced the round without a size or valuation.

Papadopoulos served as CEO for nine years. His move to chief product officer gives Llado responsibility for turning a research-driven data platform into a commercial layer for AI agents, with regulated industries serving as the initial market.

Llado spent nearly 26 years in technology leadership at Merck and later served as chief information officer at Alexion Pharmaceuticals. He joined TileDB's board in 2024. That history puts a former buyer of enterprise technology in charge of selling Tile.ai to organizations where access controls, audit trails and reproducible results are operating requirements rather than procurement extras.

A research project follows its customers into AI

Papadopoulos started TileDB as a research project in late 2014 while working as a senior research scientist at Intel Labs and participating in the Intel Science and Technology Center for Big Data at MIT CSAIL.

He spun it out in 2017 around a technical thesis: multidimensional arrays could provide a common storage model for data that did not fit cleanly into relational tables.

That work produced the open-source TileDB array engine and a 2017 VLDB paper. Papadopoulos' background also included a Ph.D. in computer science from the Hong Kong University of Science and Technology, where his dissertation focused on authenticated query processing. His research covered database security, privacy and verifiable computation.

The new governed-data pitch brings those concerns back to the center of the product. Papadopoulos said in the announcement that TileDB customers began asking over the past two years how to expose complex data to AI while keeping it "where it already lives" and controlling who could access it.

Tile.ai says the Enterprise AI Data Substrate connects data held in files, instruments, databases, software applications and cloud services to an organization's existing AI tools. The stated goal is to let agents discover and use data in place while retaining access policies, lineage, auditability and reproducibility.

Those capabilities remain Tile.ai's claims. The product is in private preview, and Tile.ai has not published performance results, deployment figures or named design partners for the new offering.

The old engine keeps its name

The rebrand does not retire the technology Papadopoulos spent more than a decade developing. Tile.ai said it will continue maintaining the open-source array engine under the TileDB name.

That separation gives the TileDB project continuity while allowing Tile.ai to market a broader commercial product. Tile.ai positions the substrate as a layer alongside existing databases, warehouses, lakehouses and catalogs. It is meant to connect those systems to agents while leaving data and established controls in place.

This is a narrower commercial message than asking enterprises to adopt a universal database as the center of their data architecture. TileDB used that framing when it raised a $34 million Series B in October 2023. AlleyCorp led that round, with Two Bear Capital, Amgen Ventures, Nexus Venture Partners, Intel Capital and several strategic investors participating.

Tile.ai now enters a crowded market in which large platforms already sell their own governance layers. Databricks offers Unity Catalog across its platform, while Snowflake has continued expanding from data warehousing into governance and AI workloads. Smaller vendors are also building control planes, connectors and data-preparation systems around enterprise agents.

Tile.ai's intended distinction is its ability to work across systems and data types without requiring another migration. Private-preview deployments will have to establish how consistently that approach works across identity providers, models, databases and on-premises infrastructure.

Llado takes on the commercial test

Llado's appointment divides the work at a consequential point. Papadopoulos retains control of product and technical direction, while Llado takes ownership of commercial scale and deployment inside the regulated organizations Tile.ai is pursuing.

In a 2024 interview with TileDB, Llado described overseeing technology through a period of substantial growth at Alexion before AstraZeneca acquired the drugmaker in 2021. His experience gives Tile.ai direct knowledge of how life-sciences CIOs evaluate data access, infrastructure risk and compliance.

The investor lineup reinforces that focus. Two Bear Capital increased its existing investment, and Amgen Ventures returned after participating in earlier TileDB rounds. The announcement gives no way to measure the additional runway, leaving the backers' identities as the clearest information about the financing.

Papadopoulos is betting that TileDB's array engine and years of work with complex scientific data can support a wider control layer for enterprise agents. Llado must turn that technical foundation into deployments that survive security reviews, integrate with entrenched data platforms and produce measurable results for buyers. The founder has moved back to building the product; a former customer-side CIO now has to prove enterprises will buy it.