Can your agents actually ship? Compose your AI-native SDLC

Can Your Agents Actually Ship? AI-Native SDLC Meetup in New York

For an AI agent to move beyond code suggestions, it needs data it can identify and interpret—and a delivery process that can check what it produces. VirtusLab’s Nov. 2 New York meetup puts those challenges together, opening with data architecture before turning to an AI-native software development life cycle.

First, make the data usable

Grzegorz Król’s opening talk, “From a Data Swamp to a Data Agent You Can Trust,” starts with a practical problem: an AI agent can return the wrong result when information is difficult to locate, unclear in meaning or restricted by permissions it cannot interpret. Król is a Tech Lead and Data Architect at VirtusLab; his work, as described in the listing, centers on letting business needs shape data architectures, models and pipelines.

The talk organizes that challenge around four capabilities—finding, understanding, securing and trusting data—and the ten technical layers the organizers say support them. Its advertised approach draws on open standards and patterns from production environments rather than a single vendor framework. Attendees are promised a diagnostic for identifying which layers an organization already has, where gaps remain and what practical first steps might look like without a large initial investment.

Further reading: Can your agents actually ship? · Luma ↗ · From a Data Swamp to a Data Agent You Can Trust · Luma ↗

A shared theme, with one agenda detail unresolved

Luma names Artur Skowroński, VirtusLab’s Head of Application Development, among the speakers. His listed experience spans more than a decade across software engineering, technical leadership, architecture and technical product management; he also writes a weekly roundup of the JVM ecosystem. The listing does not assign him a particular segment. Król is specifically identified as the opening-talk speaker.

The agenda descriptions differ on the main session’s structure. Meetup lists a 7 p.m. panel about real stories from AI adoption, followed by a separate Visdom demonstration at 7:45. Luma presents a single main session beginning at 7, with demonstrations and Q&A, rather than those separately timed segments. Both listings put the closing networking period at 8:15–9 p.m. If the panel-versus-demo distinction affects when you plan to arrive or leave, check with the host.

Further reading: Can your agents actually ship? Compose your AI-native SDLC, Mon, Nov 2, 2026, 6:00 PM | MeetupBending Spoons ↗ · Can your agents actually ship? · Luma ↗