ELC Annual 2026: AI Speed Meets Engineering Accountability

ELC Annual 2026 is scheduled for October 13–14 in San Francisco, bringing Engineering Leadership Community's conference to Fort Mason Center Gateway Pavilion. Its published program gives engineering leaders a concrete set of questions about AI: when agents can act independently, how people review the resulting work, and how the business measures the return.

For CTOs, engineering executives, and managers, those questions meet at the point where somebody becomes accountable. A workflow can produce another change before its reviewer understands the previous one. An agent can carry out an instruction before its permissions have received the scrutiny that a human decision would attract. ELC's agenda offers a useful way to examine those handoffs before they become someone else's emergency.

The gathering remains upcoming. Its relevance comes from the operating problems the organizers have chosen to discuss, with scheduled participants from organizations including 1Password, Okta, Anthropic, Hinge Health, and Slack. The prospective value is a clearer management decision to take home, rather than a promise that attending will deliver a particular financial outcome.

Agent permission becomes a leadership decision

On October 13, Rashi Agrawal, Head of Agentic AI and ML at Hinge Health, is scheduled to host a roundtable on agents acting without human approval. The session description addresses readiness evidence, confidence measures, and fallback mechanisms. It puts a practical boundary around an expansive technology conversation: the moment a system receives authority to do something consequential.

An engineering leader considering that boundary needs to distinguish tasks as well as capabilities. Permission to draft a response differs from permission to send it; access to an account differs from authority to change it. These are illustrative operating distinctions, and the conference description does not prescribe a universal threshold for every deployment. The relevant judgment depends on the action, its consequences, and the available path back when something goes wrong.

That makes the approval conversation useful to founders as well as managers. A product demonstration shows what a system can do under the conditions presented. A production decision requires someone to specify the conditions under which it can continue, stop, or ask for help. At ELC, the advertised discussion creates space to compare how other operators think about those decisions, while responsibility for each deployment stays with the organization making it.

Code review has a human capacity limit

David Harrington of Slack is scheduled to host an October 13 roundtable about code review, speed, safety, and AI. The organizers describe pressure on review as teams produce more AI-assisted changes. Their framing places attention, ownership, security, and correctness inside the same discussion, making the review process part of the adoption decision.

The management consequence is easy to miss when the visible result is another completed task. A change still has to fit the surrounding system, be understood by somebody who can maintain it, and reach a customer without creating a surprise elsewhere. Generating the change and accepting responsibility for it are separate pieces of work. This is an operating interpretation of the session's topic, rather than a measured claim about every team using AI.

For a leader attending ELC, the useful preparation is a precise account of where review currently stalls. That might involve changes waiting for specialist attention, unclear ownership, or a mismatch between what a tool produces and what a reviewer needs to evaluate. Bringing that question gives a peer discussion something concrete to work on. It also gives the leader a way to assess whether a proposed practice addresses the team's actual constraint.

AI spending needs a business explanation

The October 14 session on AI investment, risk, and delivery lists Nancy Wang of 1Password, Monica Bajaj of Okta, and Katelyn Lesse of Anthropic. The program connects security exposure, infrastructure and token costs, software quality, and customer value. It asks what leaders should measure when deciding to fund, expand, or stop an AI investment. DevCuration’s coverage of Jellyfish’s engineering cost and adoption briefing offers a related reading path for the measurement question.

Nancy Wang's role as CTO is corroborated by 1Password's appointment announcement. Katelyn Lesse is listed as Head of Platform Engineering at Anthropic. Those affiliations give the session a relevant mix of identity security, enterprise software, and AI platform operations. They establish who is scheduled to discuss the problem, while the discussion's eventual conclusions remain ahead of the event.

For the engineering executive, a useful investment explanation has to connect an operating change to something the business can recognize. Where did the released capacity go? What happened to the cost of delivering and maintaining the work? Which risks now require attention from another team? A faster development step can be valuable, but the budget conversation continues through the rest of the workflow. ELC's announced session gives attendees a reason to examine that longer chain before expanding a pilot.

Peer comparison starts with a specific problem

Engineering Leadership Community advertises curated 1:1 meetings, peer groups, and roundtables alongside the conference's speaking program. These formats are relevant because leadership choices arrive with context: team size, customer obligations, access controls, and the people available to support the work. Comparing that context can make a conversation more useful than collecting another broad recommendation. DevCuration’s earlier coverage of ELC and LinkedIn’s New York engineering roundtables provides a related leadership discussion to explore.

For a manager, the practical aim might be to understand how a peer assigns review ownership. For a CTO, it might be to examine the evidence behind a delegation boundary or an investment request. The announced program supports those questions, although any networking or learning benefit depends on the conversations that actually take place. Attendees still need to judge whether advice from another organization transfers to their own.

The physical venue is Fort Mason Center's Gateway Pavilion at 2 Marina Blvd in San Francisco. ELC's official page lists a $895 Standard Ticket for the 2-day event and a check-in window of 8:30–9:30 AM PDT. Readers considering attendance should consult the official registration page for current terms and program details. For the leaders heading into October 13–14, the most consequential conversation may begin with the approval, review queue, or budget decision already waiting on their desks.