Building Autonomous Agents at Scale

Description

Assigning a task to an agent is straightforward. Allowing it to function within a real engineering organization is more complex.Before agents can operate independently, teams must address several practical questions. What resources can the agent access? How can you manage token usage effectively? And how can you provide an entire engineering team with a unified setup without causing security issues?This meetup gathers three engineers who are currently tackling these challenges. We will learn about securing agent credentials, minimizing token consumption, and how to develop agents for scalability.For engineers and technical leads who are creating with coding agents. Free, but with limited seats.Agenda* 17:30–18:00 Welcome* 18:00–18:20 Talk 1* 18:20–18:40 Talk 2* 18:40-19:00 Break* 19:00–19:20 Talk 3* 19:20–20:00 CheersTalksTalk 1: Stop Paying Your Agents to ForgetYahav Ohana (Co-Founder Boost @JFrog)Agent loops can quickly deplete tokens by repeatedly accessing the same files, carrying unnecessary context, and retrying failed attempts.Yahav will share practical strategies to reduce token usage while maintaining agent effectiveness.Talk 2: Don't Give Your Agent the KeysPablo Goldszeft (Principal Engineer @ Islo, previously at Snyk)Agents require access to code, infrastructure, and internal tools to perform meaningful work. Providing them with broad, permanent credentials is the quickest path to a security incident.Pablo will discuss how to grant agents the necessary access without relinquishing complete control.Talk 3: The First Step Toward Autonomy is a Shared RepositoryTal Aharon (Engineer @ Legit Security)At Legit Security, a single repository grants 60 engineers access to the same agent toolchain: 27 plugins, 117 skills, and 11 MCP servers.Tal will explain how they established this system and why guardrails do not inhibit autonomy. Instead, they foster the trust needed to grant agents greater access, responsibility, and time to operate independently.