Gradient Descending with Earlybird VC: Securing AI Agents

​Gradient Descending is a series of curated roundtables on AI, bringing together technical founders and engineers for deep dives into the topics shaping how AI is built. Past discussions have covered RL Environments, Evals, Fine-Tuning, Agent Frameworks, Open Source Models and Knowledge Graphs, with speakers from OpenAI, Anthropic, Cursor, Scale AI and many more.

For this session, we’re excited to welcome Luca Palmieri, CTO and co-founder of one of Italy’s most exciting new AI companies. Currently still in stealth, Luca and his co-founder are building in the agent security space and have just closed a major Seed round, which is yet to be publicly announced. So while we can’t tell you everything just yet, this will be a rare opportunity to hear directly from the team about the technical problems they’re tackling before the company properly comes out of stealth.

​As AI agents move into production and gain access to tools, data and real-world systems, security becomes a very different problem. We’ll discuss what actually needs to be secured when agents can reason and act autonomously, the new attack surfaces this creates and what engineering teams need to think about when building agentic systems for the real world.

​Luca will kick things off with a short technical introduction, followed by an open, Q&A-led roundtable. As always, we’re keeping the room small and highly curated: up to 30 technical founders, CTOs and engineers.

​Agenda:

• ​Arrivals: 16:30–17:00

• ​Roundtable discussion: 17:00–18:00

• ​Drinks in the garden: 18:00–19:30

​See past roundtables here:

​https://www.akashbajwa.co/t/ai-roundtables

​https://ubelluzzo.substack.com/

​I agree that photographs and/or video may be taken during the event and used by the hosts on social media and in future promotional materials.