Signals Berlin 2026

Reliability in the Age of AI

Berlin · September 10–11, 2026

Signals Berlin is a curated, single-track conference for practitioners exploring how AI is changing software engineering, operations, and reliability.

We are deliberately keeping both the program and the audience small. The goal is not to maximize stage time. It is to create a high-quality conversation among people who are actively shaping this field.

AI, and especially agentic coding, are changing how systems are built, changed, and operated. Software is becoming dramatically more malleable. Small teams can now ship end-to-end systems with unprecedented speed. This creates enormous opportunity — but it also introduces new forms of fragility, new operational risks, and new reliability challenges.

This shift matters for reliability engineering in at least three important ways:

1. AI is changing how software is built.
   AI-assisted development is reshaping software engineering itself, affecting how systems are designed, implemented, reviewed, and maintained.
2. AI is changing software operations.
   AI is becoming a powerful accelerator in production, supporting debugging, instrumentation, analysis, triage, and operational workflows.
3. AI-powered systems bring new failure modes.
   As AI-heavy components become part of production software, they introduce new observability challenges, new operational complexity, and new patterns of failure.

We explicitly invite submissions from practitioners in site reliability engineering, platform engineering, DevOps, MLOps, data engineering, and related fields. Signals is meant to bring these communities into the same room and create meaningful exchange across their boundaries.

We are looking for practitioner talks, field reports, emerging patterns, conceptual models, and honest lessons from the front lines. We welcome work in progress, unresolved questions, and ideas that are still taking shape, as long as they are grounded in real practice.

There will be a separate lightning talk CFP for ticket holders closer to the event.

Topics

We are especially interested in submissions that help us understand the operational consequences of this shift, including topics such as:

• AI and software delivery
  • Coding, review, testing, and CI/CD
  • Risks from AI-generated code: defects, drift, and maintenance burden
  • Evaluation, verification, and trust in AI-assisted workflows
  • Ownership, accountability, and the changing role of the software engineer
• AI in software operations
  • Telemetry, alerting, triage, and incident response
  • Operational copilots: where they help, and where they create false confidence
  • On-call support, postmortems, runbooks, and operational knowledge systems
• Reliability of AI systems
  • Operating LLM-powered products, agents, and model-driven systems
  • Observability, debugging, retrieval, evaluation, and feedback loops
  • Uncertainty, non-determinism, data quality, and human oversight
• Platform and organizational change
  • Platform support for agentic workflows
  • Interfaces between SRE, platform, data, software engineering, and MLOps
  • Team structure, operating models, skill profiles, and engineering culture

What we are looking for

We are looking for talks grounded in real operational experience, with clear lessons, tensions, or tradeoffs, and honest reflection on what worked, failed, or remains unresolved.

We are less interested in generic product AI talks, sales pitches, broad “AI changes everything” narratives without operational depth, or abstract model research with no clear production consequences.

Format and selection

Submissions will be reviewed by the Program Committee and selected for both individual quality and overall fit. Good submissions may still not be selected if they overlap with other talks or do not fit the final shape of the program.

Timeline

• CFP closes: Sunday May 10, 2026 – 6pm CEST
• Applicants notified: May 22, 2026
• Program announced: May 28, 2026

Last updated: 2026-03-25 10:33 CET