Senior engineers in 2026 are being asked to design solutions for the failure modes of autonomous agents, keep p99 latency predictable during traffic spikes, re-architect APIs that have drifted from their original purpose, and keep legacy systems running smoothly. The 12 tracks for QCon San Francisco 2026, published this week, are organized around exactly that mix of work.
Who's selecting the program
The program committee consists of six senior practitioners, each of whom has shipped the kinds of systems the tracks will cover:
- Erin Doyle, Founding Engineer @Quotient and Instructor @Egghead
- Haley Tucker, Retired Principal Software Engineer for Platform Engineering @Netflix
- Hien Luu, Senior Engineering Manager @Zoox and author of MLOps with Ray
- Khawaja Shams, Co-Founder & CEO @Momento (previously @NASA and @Amazon)
- Suhail Patel, Staff Engineer @Monzo, previously @Citymapper
- Thomas Betts, Senior Laureate Software Architect @Blackbaud
Sessions are reviewed for commercial intent before selection, and talks that look like product roadmaps dressed up as editorial don't make it through. QCon is created for practitioners, by practitioners.
The conference runs November 16-18 at the Hyatt Regency San Francisco, with training days on November 19-20, and 60+ practitioners speaking over the three conference days.
Production AI and data engineering
The four AI tracks focus on what changes when AI has to live in production: agent architectures with real blast radius, how to evaluate systems whose behavior isn't deterministic, safety nets that have to work when the model is confidently wrong, and data platforms that serve analytics and AI workloads off the same infrastructure.
Architecting for Agents addresses software architecture as autonomous agents become first-class actors. Topics explored include orchestration, boundaries, control, and emergent behavior.
Engineering AI Systems goes beyond prototypes and into production, covering the realities of building, scaling, and maintaining AI-driven systems in the real world.
Guardrails & Safety Nets (Evals) for the New Landscape looks at how teams are building evaluation systems, feedback loops, and safety mechanisms to make AI-powered systems reliable and trustworthy.
Data Platforms Reimagined examines how modern data platforms are evolving to support real-time, AI-driven, and decentralized use cases, and how that is redefining the way data flows through organizations.
Architecture and resilience
Three tracks cover systems problems that aren’t focused on in the current hype cycle. These include architecture teardowns of systems that senior engineers have only heard about, distributed systems behavior under real load, and designing for failure instead of hoping against it.
Architectures You've Always Wondered About covers systems people have heard about but never fully understood, and deep dives into real architectures, trade-offs, and the decisions that shaped them.
Something About Distributed Systems shares real-world lessons of building and operating distributed systems in production, and focuses on topics such as latency, consistency, failure, and scale.
Resilience Engineering: Everything Fails All the Time examines how to design systems that anticipate failure, adapt to it, and continue operating under stress.
Platforms and developer experience
Three tracks are about the internal systems engineers' ship for other engineers. Platforms can either make other teams faster or become another approval gate. APIs can either survive a decade of requirement changes or become the next legacy migration. Developer experience is either architecture now or a Slack channel full of complaints.
Real World Platform Engineering looks at how organizations are building internal platforms that actually work, balancing standardization with flexibility to accelerate developer productivity.
Developer Experience as a First-Class Architecture Concern treats developer experience as a core architectural problem, exploring how tooling, workflows, and system design directly impact productivity and outcomes.
Modern API Design is about designing APIs that stand the test of time, focusing on usability, evolvability, and real-world constraints.
People and practice
Two tracks step back to the people building the systems: what operating at Staff+ actually requires when the work is as political as it is technical, and what happens to engineering roles when coding moves outside of engineering entirely.
Staff+ Engineering Skills covers what it really takes to operate at Staff+ level. From influencing without authority to driving technical strategy across teams and organizations.
Code Beyond Engineers & Engineering Teams looks at how coding is expanding beyond traditional engineers into product, operations, and AI-assisted roles, and what this means for teams and systems.
Speakers are being announced on a rolling basis. Conference tickets are $2,355, with early bird pricing and team discounts available until May 12th. Register at qconsf.com.