From Startup to AI-First: Braze CTO Jon Hyman on Engineering Transformation

Introduction

When Jon Hyman co-founded Braze nearly 15 years ago, the engineering landscape looked vastly different. Today, as CTO of the customer engagement platform, he oversees a team that has undergone a radical shift—transforming from a traditional engineering organization into an AI-first powerhouse in just a few months. In this article, we explore how Hyman rethought engineering for the agentic era, scaling growth while embracing artificial intelligence as the core driver of product innovation.

From Startup to AI-First: Braze CTO Jon Hyman on Engineering Transformation — Source: stackoverflow.blog

The Challenge of Scaling Engineering Over a Decade

Braze started as a small startup with a handful of engineers. As the company grew into a publicly traded firm with thousands of customers, Hyman faced the classic scaling conundrum: how to maintain speed and quality while expanding the team. Over nearly 15 years, he learned that process must evolve with size. Early on, feature velocity was king. But as the engineering org ballooned, Hyman and his leadership introduced specialized units—platform, product, and infrastructure teams—each with clear ownership.

Key Milestones in Braze's Engineering Growth

Year 1-3: Monolithic architecture, all-hands-on-deck approach
Year 4-6: Microservices adoption, formalized code reviews
Year 7-10: SRE culture, incident management, and data-driven decision-making
Year 11-15: Shifting to AI-first, embedding ML engineers across product squads

The AI-First Pivot: A Matter of Months

In late 2023, Hyman and his team recognized that the agentic era—where AI agents could autonomously perform tasks—required a fundamental rethinking of how engineering operated. Instead of treating AI as a separate feature, they decided to embed it into the very fabric of development. Within three months, they restructured into cross-functional squads that included data scientists, ML engineers, and product managers working side by side.

What Is the Agentic Era?

The agentic era refers to a paradigm where AI systems not only generate content or predictions but also take actions—like personalizing customer journeys or autonomously optimizing marketing campaigns. For Braze, this meant building agents that could learn from user behavior and act without human intervention. Hyman notes that the shift required not just new tools but a new mindset: trust in AI's autonomy was the biggest cultural hurdle.

How They Transformed Engineering in 90 Days

The transformation was not a slow migration. Hyman's playbook included three key actions:

Refocus hiring: Prioritized candidates with both software engineering and machine learning experience. Existing engineers were also upskilled through internal bootcamps.
Adopt AI-native tools: The team switched to using AI-assisted code generation, automated testing, and intelligent monitoring systems. These tools reduced toil and freed engineers for higher-level problem solving.
Reorganize around outcomes: Instead of teams owning components (like "API team" or "Data team"), they created mission-oriented squads named after customer outcomes, such as "Campaign Optimization" and "Predictive Targeting."

The Role of Leadership

Hyman emphasizes that the CTO's job during such a shift is to set the vision and remove blockers. He personally led weekly AI town halls to communicate progress and answer questions. "Engineers are naturally skeptical," he says. "You need to show them the why before they invest in the what."

Results and Lessons Learned

Early results from Braze's AI-first transformation include a 40% reduction in manual configuration time for marketers and a notable improvement in customer engagement metrics. Hyman also learned that infrastructure costs initially rose as they trained models, but efficiency gains quickly balanced the books.

Key Takeaways for Engineering Leaders

Move fast but transparently: The 90-day timeline was aggressive, but weekly updates kept anxiety low.
Invest in AI education: Not every engineer needs to be a data scientist, but everyone should understand the basics of AI model lifecycle.
Measure what matters: Hyman focused on deployment frequency and model accuracy instead of lines of code or story points.

Looking Ahead: The Next 5 Years

Hyman believes that the agentic era will only accelerate. Braze is exploring autonomous experimentation—where AI runs A/B tests without human input—and self-healing infrastructure that predicts outages. For the engineering team, the journey from a startup to an AI-first organization is ongoing, but the foundation laid in those few months has positioned them for the next decade of growth.

Jon Hyman's story is a testament that even long-tenured engineering cultures can pivot to AI-first leadership with the right combination of vision, speed, and trust.

Tags: