Register and pick up your badge day prior to event so you can skip lines next morning.

After collecting your badge, kickstart your day with a delicious complimentary breakfast and freshly brewed coffee.

The opening ceremony of Shift Miami 2025 marks the beginning of an inspiring journey! This session sets the stage by reflecting on our progress and embracing the incredible opportunities ahead. As AI and developers come together to shape the future of software, join us to explore groundbreaking insights, hands-on innovations, and the transformative power of AI in development.

Planning, coding, testing, monitoring—the endless cycle that spans 10+ tools that fragment our focus and slows delivery to a crawl. Vibe coding doesn't work when you've got millions of lines of code. But what if you could orchestrate multiple AI agents on tasks beyond just generating code, while you focus on the creative decisions that matter? In this talk, I'll demonstrate how real enterprise organizations are changing their entire SDLC—going from understanding, planning, coding, and testing all the way to incident response—using AI agents.

TBA

Remember when you started working on that application and only admins could do that one action? Then that quickly evolved to having others being able to do that other action and you ended up with a bunch of different roles, attributes, etc to handle all those cases? Ha, welcome to the world of access control systems! In this talk, we’ll follow the journey of a project that begins with no access control, progresses to Role-Based Access Control (RBAC), struggles with Attribute-Based Access Control (ABAC), and ultimately finds its footing with Fine-Grained Authorization (FGA). You'll see how this journey becomes even more important in modern applications, especially when thinking about things like AI. In this process, you’ll learn how OpenFGA addresses the growing complexity with a relationship-based model that’s both flexible and scalable.

Move beyond text-based paradigms and explore the frontier of Artificial Intelligence: multimodal applications that see, hear, and interact with the world in real-time. This technical keynote explores the practical implementation and potential of systems integrating and synthesizing information from diverse data streams - including live input from webcams, audio feeds, video, mobile sensors, and bespoke hardware.

Stefania Druga (Google DeepMind) will provide a researcher's perspective, diving into the technical challenges - from data fusion and latency reduction to context modeling and robust interaction design - inherent in building AI that leverages a richer understanding of the physical world. Expect compelling live demonstrations showcasing interactive AI systems designed to understand context, anticipate user needs, and respond dynamically through multi-sensory feedback loops.

Discover the future trajectory of intelligent systems and the potential for innovation in human-computer interaction, personalized assistance, and human-AI collaboration enabled by this shift towards truly multimodal AI

TBA

Legacy code is everywhere, from swiping your credit card to booking your next flight. Behind the scenes, critical systems running on decades-old code (like COBOL on mainframes) power essential services. These systems, evolving over 40+ years, form the backbone of many businesses. In this talk, I'll share our journey and how we approach modernising these critical legacy systems.

TBA

Refuel, relax, and get ready for more great sessions ahead!

Instead of relying on a single agent rollout, people should embrace the idea that having multiple agents tackle the same problem and try to solve them independently with a strong verification loop leads to better answers. In a world where compute is cheap, this allows us to further scale up the performance of any given agent.

TBA

AI-driven development is helping software engineers build and ship code faster than ever. But some engineers are finding the transition easier than others--many still haven't adopted autocomplete, let alone fully autonomous agents. And major problems arise when we don't use this technology properly.

In this talk, we'll cover best practices for using software agents to fix bugs, implement features, and reduce the toil involved in day-to-day software engineering. We'll also go through some real-world examples with OpenHands, an open source software development agent.

Have you always been interested in robotics or fun "maker space" type inventions, but had no idea where to start? That was Adrienne; she longed to see if she could actually build some of her crazy ideas, but was too intimidated and overwhelmed to begin. Luckily, her role at Viam has given her the extra push she needed to finally start.

Join Adrienne as she takes you through her learnings of what robots are, how they work, and how to get started with robotics without knowing too much about hardware or robots at all!

AI coding assistants promise productivity but hide a major performance tax. Our research reveals a shocking truth: 90% of AI-suggested code optimizations either break functionality or fail to improve performance.

In this eye-opening session, we'll demonstrate:

- Why AI code generation systems consistently generate inefficient code
- A verification system that detects and fixes these issues before they reach production
- How we can automate code optimization for maintaining AI productivity without sacrificing performance
- Real-world examples where automated optimization sped up an entire computer vision pipeline

Join Saurabh Misra (Founder, Codeflash) as we expose the hidden costs of AI-assisted development and provide actionable solutions to always ship performant code.

Refresh with a cup of coffee and a quick chat!

Between theories on eventual impact of AGI and most basic of the LLM use cases (e-mail anyone?) - most product teams fail to utilize AI.

This talk dives deep into actual use-cases and lessons learned at applying AI (not just LLMs) to Product Discovery, Research, Data, Shipping, and Evaluation at Kiwi.com.

Goal is to use AI as a force multiplier for productivity, instead of a crutch, enabling Product teams to do more with less, while building products that are better for both the users, and the business.

Generative AI is used every day by millions of developers and vibe coders to generate functions, data processing pipelines, games or entire applications and, realistically, a lot of that code sucks to the point that it sometimes won't even run. The piece that AI Agents need to generate runnable code is access to secure, fast and scalable remote sandboxes.

In this talk, we explore how access to such tools opens up possibilities for, not only running generated code, but for highly parallel workflows such as data analysis or reinforcement learning. We'll explore the technical requirements to build sandboxing tools and how to connect Agents to existing solutions.

TBA

Wrap up the day with a short closing ceremony.

TBA

Connect with fellow participants to get ready for the upcoming workshops at The Lab Miami.

TBA

The cloud is no longer a one-size-fits-all solution. As organizations strive to stay competitive, multicloud strategies are becoming essential for leveraging the strengths of multiple cloud providers, enhancing resiliency, and optimizing costs. But how can developers efficiently manage infrastructure across AWS, Azure, Google Cloud, and others—without drowning in complexity? And on top of that how do we utilize AI efficiently in our stack?

In this talk, we’ll show you how to harness the power of Infrastructure as Code (IaC) to build, deploy, and scale AI multicloud applications with confidence. From choosing the right IaC tools to crafting workflows that eliminate bottlenecks, this session is packed with practical insights and real-world examples to help you succeed in a multicloud world.

We all need a little boost, don't we?

The rise of AI-driven applications has made vector search a game-changer for tasks like recommendation systems, semantic search, and Large Language Model (LLM) enhancements. But how do you get started with vectors, similarity search, and LLMs using open-source tools?

In this hands-on workshop, you'll learn the fundamentals of vector embeddings, how similarity search works, and how to integrate it with LLMs for powerful AI applications. We'll explore key open-source tools like FAISS, Apache Pinot, Milvus, and OpenAI's embeddings API. Whether you're a developer, data scientist, or AI enthusiast, this workshop will provide the building blocks you need to leverage vectors in real-world applications.

TBA