The Video Signal

Topic

Nubank

All digests tagged Nubank

· 16:24

We Vetted 2000 AI Skills Before They Reached Developers — Lucas Palma, Nubank

Lucas Palma details how Nubank addressed AI skills—which function like supply chain dependencies—being a significant security risk in a regulated financial environment. To mitigate this, they built 'Skill Vector,' a comprehensive security review system that sits between the skill and the internal marketplace. This system vets every skill using a hybrid approach: deterministic checks (e.g., regex for unsafe instructions) followed by LLM analysis to assess context and behavior before granting permissions or allowing deployment.

Key takeaways

  1. Treat AI Skills as Dependencies

    AI skills must be treated like any other software dependency, requiring rigorous vetting and security gates, rather than just configuration files.

  2. Hybrid Security Scanning is Essential 12:17

    The most effective approach combines deterministic checks (for easy risks) with LLM review (for contextual behavior analysis), ensuring comprehensive coverage.

  3. Security Gates Must Be Comprehensive

    Vetting must cover the entire workflow, including local developer testing, CI/CD steps, PR feedback, and integration into a formal vulnerability management program.

Watch on YouTube Full article

· 17:38

How Forward Deployed Engineering is done at Cognition — Jia Wu

Forward Deployed Engineering (FDE) at Cognition focuses on maximizing the overlap between product capabilities and complex customer problem spaces, moving beyond simple code generation. The core value proposition is measured by tangible business outcomes—such as reducing delivery timelines or increasing PR acceptance rates—rather than merely token usage. FDE embeds within client ecosystems to identify high-leverage areas for AI agents like Devin, addressing the full software development lifecycle (SDLC) challenges including testing, review, and maintenance.

Key takeaways

  1. Shifting Focus from Tokens to Outcomes

    The industry trend is shifting away from optimizing for token usage toward measuring true business outcomes. Cognition measures value by demonstrating a real delta—for example, an 82% reduction in delivery timelines after agent activation.

  2. FDE's Role: Maximizing Problem-Product Overlap 8:36

    FDE aims to maximize the intersection between existing product domains and critical customer problems. This requires deep understanding of the client’s problem space, identifying the highest leverage points for automation.

  3. The Value is Non-Linear 3:25

    Adoption of AI agents follows a non-linear curve: initial use may be a step function, but enterprise-wide adoption and integration lead to parabolic growth as product capabilities align with the customer's full backlog.

  4. Core Mandates of FDE

    FDE operates under two mantras: relentlessly tying into the customer (who are the 'lifeblood') and prioritizing correctness and customer success at all costs. The goal is to make the customer successful, not just ship code.

Watch on YouTube Full article