ELC Annual 2025

Harness is proving that innovation at scale doesn’t have to come at the cost of speed or focus. By adopting a “startup within a startup” model, independent teams operate with the agility of a startup while leveraging shared infrastructure to move faster and smarter. This approach is blurring the lines between engineering, product, and business, requiring leaders to build cross-disciplinary skills, create tighter alignment on outcomes, and embed accountability at every level. In an era where AI is accelerating both opportunity and competition, this model offers a blueprint for scaling innovation that directly drives business impact.

In this session, you'll get insights into:

• How Harness structures independent startup teams to innovate quickly while maintaining accountability
• Essential product and business skills engineering leaders need to become true strategic partners
• Strategies for aligning roadmaps, investment decisions, and metrics across engineering and product
• How to connect engineering work directly to business outcomes to maximize impact
• Leadership principles for fostering rapid innovation cycles that scale without losing focus

"As AI becomes inseparable from modern software delivery, DevEx leaders are at a crossroads. No longer just stewards of tools and workflows, they are uniquely positioned to lead the AI transformation across engineering organizations.

In this session, leaders from Expedia Group and LinearB will make the case for why DevEx must own AI strategy—not simply support it. With AI reshaping code generation, review, testing, and release, the next frontier of developer productivity depends on how well DevEx operationalizes AI. Attendees will learn why measurement without action falls short, and how AI—when harnessed by DevEx—can turn data into decisions and insights into automation. You’ll leave with a bold new lens on AI’s role in engineering and why DevEx is poised to lead the way."

Engineering leaders are facing a new challenge: proving AI is actually delivering impact. In this session, Jellyfish shares its AI Impact Framework – built in collaboration with 300+ engineering leaders and grounded in data from over 20,000 developers. Learn how to assess adoption, productivity, and business outcomes with a structure tailored to today’s hybrid AI-human development teams.

This is a practical look at building AI agents that drive business impact with Lin Qiao, CEO of Fireworks AI and co-creator of PyTorch. She’ll share strategic frameworks for agent design (from product-model co-design to sustainable data flywheels), plus case studies from customers Cursor, Uber and UiPath. Learn how top engineering teams approach model selection, fine-tuning, and scaling to outperform closed-source solutions. This session covers the full agent lifecycle, with insights for evaluating agent opportunities, deploying to production, and adapting to the fast pace of model evolution.

Often companies struggle with the role of engineering leadership as they evolve and grow. Some try to avoid formal management altogether for too long and then overreact adding layers of leaders who revisit and slow decision making. As AI changes software engineering, teams will become smaller, more focused, and efficient. Leadership needs to meet the challenges of the new environment, rather than clinging to old tropes. There are not set answers today but we need to lay out the challenges and the possible solutions.

Join Bill Coughran, Partner at Sequoia Capital and former SVP of Engineering at Google for a fireside chat to discuss the challenges and opportunities ahead.

In less than a decade, OpenAI evolved from a pure research lab into the fastest-growing product in history, scaling from 100 million to 700 million weekly users. This transformation required building an engineering function and culture from scratch, blurring the boundaries between research, product, and infrastructure. Sulman Choudhry, Head of ChatGPT Engineering, shares the organizational design choices, hiring shifts, and cultural bets that enabled unprecedented shipping velocity. He’ll cover what happens when domains collide, how to balance senior talent with AI-native interns, and the real-time architecture changes behind major launches like ChatGPT Images. This is a candid look at the patterns, anti-patterns, and opportunity zones that emerge when you scale AI products at breakneck speed.

A practical guide to building AI agents that goes beyond the hype. Learn Malte's three-phase development approach, as well as his POV on agent architecture, security considerations, and technical stack requirements on the path from "works sometimes" to "fewer mistakes than humans.

Most companies approach AI adoption backwards — adding features first, culture second. At Amplitude, we flipped the script. In just five days, we transformed 200+ product builders from AI skeptics to AI-native practitioners, achieving 40% productivity gains and cutting months-long projects to weeks. This isn't another hackathon story—it's a blueprint for how leaders can create the space, pressure, and vulnerability needed to shift entire organizations from watching AI happen to making it happen. Learn the specific tactics that worked, what surprised us most, and why the real transformation started after the week ended.

The future of products will not be confined to a screen. We are entering an era where interfaces see, hear, speak, and imagine alongside us: powered by multimodal AI. These systems go beyond clicking and tapping, blending vision, voice, and other sensory inputs with rich, dynamic outputs like conversation, imagery, and immersive environments. The interface itself begins to dissolve, giving way to experiences that feel less like using a tool and more like collaborating with a trusted partner. In this talk, we’ll explore how to design the next generation of product interfaces where AI doesn’t just respond, it understands, anticipates, and co-creates in ways that fundamentally reshape how we interact with technology.

Traditional AI lived on simple benchmarks: accuracy, precision, BLEU scores. Generative AI broke that mold. Now, outputs are open-ended, there’s no unique gold standards, and as GenAI has been industrialized, neither datasets nor evaluation suites are shared across vendors. In this session, we’ll look at the science and strategy of evaluation in this new era: how to balance human adjudication with automated metrics, how Goodhart’s Law plays out in practice, and how evaluation itself shapes product development. Drawing on examples from healthcare, we’ll show why getting evaluation right isn’t just an academic concern, it’s the foundation for building products that customers can trust.

Everything we thought we knew about building great products is changing… or is it?

Developers and designers are staying in sync and working more efficiently than ever before with tools like Figma’s Dev Mode. At the same time, new research shows that AI is blurring the boundaries between disciplines across the team. Yet as collaboration deepens and roles broaden, domain expertise isn’t disappearing—it remains essential to building great products.

How can teams hone their respective skills while embracing a more collaborative product development process? Join Figma’s Marcel Weekes and Andrew Hogan for a conversation about what these shifts mean for product teams and how leaders can embrace a more collaborative, AI-enabled future.