AI agents need more than just language - they need to act. This talk introduces trajectory data - an emerging class of data used to training LLM agents. These are sequences of observations, actions, and outcomes that drive agent learning. Whether you're training agents or building the data pipelines behind them, this is your guide to the data powering the next generation of AI agents.

OneLake eliminates pervasive and chaotic data silos created by developers configuring their own isolated storage. OneLake provides a single, unified storage system for all developers. Unifying data across an organization and clouds becomes trivial. With the OneLake data hub, users can easily explore and discover data to reuse, manage or gain insights. With business domains, different business units can work independently in a data mesh pattern, without the overhead of maintaining separate data stores.

The data lakehouse architecture has made big waves in recent years. But there are so many considerations. Which table formats should you start with? What file formats are the most performant? With which data catalogs and query engines do you need to integrate? To be honest, it can become a bit overwhelming.

But what data engineer doesn't like a good technical challenge? This is where it sometimes becomes a philosophical decision of build vs buy.

In this presentation, Onehouse VP of Product Kyle Weller will break down the pros and cons he has seen over nearly a decade of helping organizations implement their own data lakehouses and building the Universal Data Lakehouse at Onehouse. You'll learn about:

• The strengths of open table formats such as Apache Hudi™, Apache Iceberg™ and Delta Lake
• Interoperability via abstraction layers such as Apache XTable™ (incubating)
• Lakehouse optimizations for cost and performance via Apache Spark™-based runtimes

In this talk, I will go over some of the implementation details of how we built a Data Lake for Clari by using a federated query engine built on Trino & Airflow while using Iceberg as the data storage format.
Clari is an Enterprise Revenue Orchestration Platform helping customers run their revenue cadences and helping sales teams close deals more efficiently. Being an enterprise company, we have strict legal requirements of following data governance policies.
I will cover how to design a scalable architecture for building data ingestion pipelines for bringing together data from various sources for the purposes of AI and ML.
I will also cover some of the use cases that have been unlocked in the company with respect to building Agentic Frameworks with the development of this data lake.

‍

Understanding and improving unit-level profitability at Amazon's scale is a massive challenge, one that requires flexibility, precision, and operational efficiency. It's not only about the massive amount of data we ingest and produce, but also the need to support our evergrowing businesses within Amazon. In this talk, we'll walk through how we built a scalable, configuration-driven platform called Nexus, and how Apache Hudi™ became the cornerstone of its data lake architecture.

In the coming years, use of unstructured and multi-modal data for AI workloads will grow exponentially. This talk will focus on how Ray Data effectively scales data processing for these modalities across heterogeneous architectures and is positioned to become a key component of future AI platforms

Apache Gluten™ (incubating) is an emerging open-source project in the Apache software ecosystem. It's designed to enhance the performance and scalability of data processing frameworks such as Apache Spark™. By leveraging cutting-edge technologies such as vectorized execution, columnar data formats, and advanced memory management techniques, Apache Gluten aims to deliver significant improvements in data processing speed and efficiency.

‍

The primary goal of Apache Gluten is to address the ever-growing demand for real-time data analytics and large-scale data processing. It achieves this by optimizing the execution of complex data processing tasks and reducing the overall resource consumption. As a result, organizations can process massive datasets more quickly and cost-effectively, enabling them to gain valuable insights and make data-driven decisions faster than ever before.

As the adoption of LLMs continues to expand, awareness of the risks associated with them is also increasing. It is essential to manage these risks effectively amidst the ongoing hype, technological optimism, and fear-driven narratives. This presentation will explore how to address vulnerabilities that may emerge. Our focus will extend beyond simply securing interactions with the models, emphasizing the critical role of surrounding infrastructure and monitoring practices.

‍

The talk will introduce a structured framework for developing "system-level secure" AI deployments from the ground up. This framework covers pre-deployment risks (such as poisoned models), deployment risks (including model deserialization), and online attack vectors (such as prompt injection). Drawing on two years of experience deploying AI systems in sensitive environments with strict privacy and security requirements, the talk will provide actionable strategies to help organizations build secure, resilient applications using open-source LLMs. Attendees will gain practical insights into strengthening both AI models and the supporting infrastructure, equipping them to develop robust AI solutions in an increasingly complex threat environment.

This talk delves into the revolutionary potential of AI agents, powered by generative AI and large language models (LLMs), in transforming ETL (Extract, Transform, Load) and ELT (Extract, Load, Transform) processes.
With an emphasis on automating code generation, streamlining workflows, and enhancing data quality, this session explores how AI-driven solutions are reshaping the landscape of data engineering. Attendees will learn how these intelligent agents can reduce manual coding, eliminate errors, improve operational efficiency, and meet compliance requirements, all while accelerating development timelines.
We will also cover key use cases where AI agents facilitate real-time data transformation, ensure data governance, and promote seamless deployment across cloud environments, giving businesses a competitive edge in today's data-driven world.

In today’s rush to adopt AI, many organizations overlook a critical truth: value doesn’t come from AI alone—it comes from the powerful combination of software engineering, data engineering, and AI/ML engineering. In this fast-paced, 15-minute talk, Nisha Paliwal draws on 25+ years of experience in banking and tech to unpack the "Trifacta" that fuels real transformation.

From legacy systems to self-learning platforms, she’ll share stories, stats, and insights on how this triad—when integrated—enables banks to move faster, deliver smarter experiences, and generate measurable impact.

Whether you're a technologist, leader, or change agent, you’ll walk away with a fresh lens on cross-functional collaboration, and practical ways to break silos, build trust, and unlock innovation at scale.

Two very big things happened in recent years that completely transformed the data landscape:

• Pre-trained AI models: These models have democratized AI, enabling software engineers to integrate advanced capabilities into applications with simple API calls and without extensive machine learning expertise. ​
• Lakehouses: Open formats over object storage bring together the flexibility of data lakes and the data management strengths of data warehouses, offering a more streamlined and scalable approach to data management.

And yet, most data platforms remain difficult to digest for traditional software developers.

This talk introduces "Data as Software," a practical approach to data engineering and AI that leverages the lakehouse architecture to simplify data platforms for developers. By leveraging serverless functions as a runtime and Git-based workflows in the data catalog, we can build systems that makes it exponentially simpler for data developers to apply familiar software engineering concepts to data, such as modular, reusable code, automated testing (TDD), continuous integration (CI/CD), and version control.