QCon San Francisco

These days it is possible to achieve fairly good performance on cloud provisioned systems. We discuss the design of a high performance, strongly consistent system which maintains constant service in the face of regular updates to core logic.

Platform teams frequently inherit systems that were never architected for their current scale, yet are so foundational that downtime can halt the business.

Modern AI platforms don’t have to choose between deterministic precision and probabilistic exploration—they need both.

As a college sophomore, I founded Pointz, initially building its core routing engine from scratch. I architected the platform’s safety rating algorithm, integrating geospatial data structures, crowdsourced inputs, and multi-criteria decision-making to deliver context-aware routing.

The agentic AI landscape is facing an architectural crisis due to framework fragmentation, bottlenecking the creation of truly reliable autonomous systems.

Rust runs faster, but it slows down engineers, right? This was our team’s assumption when we decided to rewrite our code from Kotlin into Rust. But we were wrong in completely unexpected ways.

AI inference is expensive, but it doesn’t have to be. In this talk, we’ll break down how to systematically drive down the cost per token across different types of AI workloads.

Have you ever wondered what it\'s like to respond to a significant incident? Walk through an hour by hour reconstruction of an incident response or two, focusing on what it was like to be \"in the room\" and the human response to the incidents.

The AI coding agent landscape evolves weekly. This talk compares today’s frontrunners, shows where each shines, and shares prompts, policies, and “rules templates” that turn code suggestions into production-quality output.

As engineering teams explore building AI agents, one challenge becomes clear: agents don’t live in isolation, and trust in agentic systems is often lacking.

What if everything you know about building distributed systems is backwards?

For decades, visualization of service metrics overwhelmingly converges to line charts. The time-centric nature of real-time telemetry further cemented this phenomenon via storage layouts and domain-specific query languages.

Maintaining up-to-date and secure software across a polyglot fleet is a challenge for any engineering organization. Manual migrations and urgent updates disrupt productivity and require coordination across many teams.

The Technical Leadership Challenge: Every successful engineering team faces the same inflection point: transforming beloved internal tools into a legitimate platform business.

Developers today are not just writing code, they’re orchestrating complex systems, managing pipelines, and maintaining reliability at scale. But when observability starts too late in the lifecycle, teams are left reacting instead of building.

Netflix runs on a complex multi-layer cloud architecture made up of thousands of services, caches, and databases. As hardware options, workload patterns, cost dynamics and the Netflix products evolve, the cost-optimal hardware and configuration for running our services is constantly changing.

As engineers, we’re used to managing failure, but long-running outages hit differently. They stretch teams, systems, and assumptions about how incidents “should” play out.

Over the past year, Zoox has invested in integrating Large Language Models (LLMs) into internal developer workflows through a company-wide initiative called Zoox Intelligence (ZI).

In high-performance code, a single misplaced counter increment can cost more than the operation it’s measuring. That creates a paradox: instrument too much and you slow the system down; instrument too little and you miss the insights you need to continuously deliver.

As enterprises seek to move AI from proof-of-concept to production, standalone vector databases face limits in synchronization, ACID compliance, and resilience. This session shows how PostgreSQL-compatible distributed databases address these issues while keeping a familiar developer experience.

Cloud systems are undergoing an architectural shift. Traditional shared-nothing designs struggle to deliver the elasticity, availability, and operational simplicity that the cloud demands.

When many teams work on a large, complex application, the microservice architecture potentially enables them to work independently and deliver a continuous stream of changes.

The Platform That Sees Risk Before Code Does

Startups are the harshest environments for engineers. Limited time, resources, and energy force teams to make decisions under pressure — and those decisions can determine whether a company survives or stalls.

AI coding tools promise productivity gains, but many teams aren’t seeing the impact.

SS&C Technologies runs 47 trillion dollars of assets on our global private cloud. We have the primitives for infrastructure as well as platforms as a service like Kubernetes, Kafka, NiFi, Databases, etc.

In this talk, we\'ll explore some of the \"ironies\" of automation—and now, artificial intelligence—in their interactions with software operators (i.e. you), especially during high consequence, high tempo situations (aka incidents).

Behind every productivity leap is a rock-solid platform. Go under the hood with Intuit’s GenOS team to see how vector stores, prompt management, RAG pipelines, and agent orchestration come together to serve ~100 million users.

Ever wondered what happens when a bunch of performance-obsessed developers decide their blazing-fast database isn\'t quite blazing-fast enough?

Details coming soon!

At Uber’s scale, ensuring an always-green mainline while processing hundreds of changes per hour is a massive challenge— especially when those changes span multiple language monorepos supporting dozens of business-critical apps.

In this talk, Matthew will describe lessons learned from over 20+ years of building scalable, secure and stable infrastructure platforms for software in financial services (electronic trading, credit card processing etc.), the talk is relevant to anyone building platforms for mission-critic

Since the first compiler, software has been a stack of human-friendly abstractions translated into machine instructions. The engineers who understood at least some of what was going on under the hood were essential—indispensable even. But now?

We are currently moving from \"AI-enabled\" systems, where artificial intelligence is an additive feature, to \"AI-native\" systems, where intelligence is the foundational, architectural core.

As AI applications demand faster and more intelligent data access, traditional caching strategies are hitting performance and reliability limits.

Routine database upgrades should be straightforward, especially with familiar, well-established technology. We were confident heading into our Elasticsearch upgrade, equipped with a solid plan and excited to see performance gains like we had seen from past upgrades.

In this session you\'ll learn how product leaders turn GenAI enthusiasm into an enterprise-ready blueprint for real productivity gains.

In order to improve the performance of existing applications and services, we can identify the most performance-critical pieces and reimplement them in Rust as opposed to completely rewriting the applications from scratch.

Most engineering talks start with buzzwords like “AI” and “personalization.” This one stMarts with a question: How do you implement in-session actions in your application?

What does it take to safely delete data at Netflix scale? In large-scale systems, data deletion cuts across infrastructure, reliability, and performance complexities.

Once a stock market darling and a pioneering hyperscaler in the 1990s and early 2000s, eBay has been in steady decline since the 2010s. A household name with a flat business, eBay has been unable to make substantive strides in its market reach or its engineering outcomes in the last 15 years.

Developing SDKs for your users in multiple languages can come at a high cost - especially if you need to implement complex logic client side, but traditionally options for sharing logic across those languages have been quite limited.

Every team wants secure container images, but few realize how complex it becomes at scale.

As companies grow, so does the complexity of keeping distributed systems in sync. At DoorDash, we tackled this challenge while building a high-throughput, domain-oriented data platform for capturing changes across hundreds of services.

In January of 2023, the Microsoft Azure Wide Area Network experienced a global outage. If you were a Microsoft customer at the time, you were impacted by this outage.

Autonomous agents operate in a continuous loop: perceive context → reason → execute tools → observe. Each edge creates distinct attack surfaces. This talk maps vulnerabilities—memory poisoning in context ingestion, goal hijacking during reasoning, blind execution at the action stage.

The popularity of Python means insurance and financial services companies have a growing body of actuaries, quantitative developers, and software engineers capable of building innovative and customized solutions for both data management and modeling.

In an increasingly competitive field, software expertise alone may not be enough to stand out and drive your career forward.

How does Netflix maintain a seamless viewing experience for millions of users, especially during traffic spikes or when backend datastores are overloaded? Autoscaling can help during traffic spikes, but it costs money, takes a few minutes to kick in, and capacity may not always be available.

Coding agents are a new tool, which many of us are trying to figure out how to use effectively.

In an early-stage startup, the lines between roles blur, and engineering can become less about building and more about what makes the company attractive.