Transforming AI experiments into enduring business value.

How Pure Math’s CEO & CTO Helped Lay the Foundation for Vetted’s Recent Acquisition by Yahoo.

Last month, Yahoo announced the acquisition of Vetted, an AI-powered product search engine where both Sean and Tom made foundational technical contributions before leaving to focus on building Pure Math AI. We recently asked them to provide some background on the path from startup to pivot to acquisition in the hope that their experience building lasting technical infrastructure to support the use of AI might offer valuable insights to firms exploring the uses and implementation of AI in their own organizations.

What they shared reveals how thoughtful architecture and team building can transform AI experiments into enduring business value.

The Foundation: Building for Reproducibility Before It Was Critical

Vetted began as Slant—a platform for subjective questions like "what's the best programming text editor" that Stack Overflow had rejected. As it evolved into a collaborative product review website where users could upvote and downvote products and statements, it inevitably became a target for brigading, flame wars, astroturfing, and SEO spam.

Tom's early decision to build the system using event sourcing—enabling complete reproducibility—would prove crucial for enabling Vetted's initial forays into data analysis.

Driven by the need to fight spam and vote brigading, this architecture allowed what would become Vetted's key transformation: evolving from a rudimentary community website into an organization with data science at the core. The moderation team's manual efforts comparing rankings against established review sites revealed patterns that would later inform sophisticated AI systems.

The Neural Network Pivot: Turning AI Experiments to Value

When TensorFlow 1.0 emerged, Slant’s leadership recognized an opportunity—these new tools could be used to automate some of the processes needed to combat spam, and Tom got to work on building a proof-of-concept. But the real transformation came when Vetted hired Sean as their first head of data science. This hiring decision marked a fundamental shift—Vetted would become a data science company at its core.

While Tom's proof-of-concept showed there was a viable path, it was Sean's experience as a data scientist that allowed him to fully rework the approach to find new signals in the data that others had missed, and deliver more precise rankings that were as stable as users had come to expect.

He developed algorithms to detect genuine deals in price fluctuations—turning noisy price data into actionable insights. He built dynamic ranking systems that adjusted to user budgets, making the platform responsive to individual needs rather than showing static lists.

While Tom and his engineering team optimized the underlying Pandas and NumPy code to handle scale, Sean was expanding what the AI could achieve with that scale. The relationship was symbiotic: infrastructure enabled data science, but data science drove the vision.

Building a Data Science Organization

The re-brand to Lustre.ai reflected this new identity. The vision was ambitious: build a fully automated ranking system by finding consensus among third-party review sites using neural networks. Sean understood that one data scientist couldn't transform a company. He built and led a growing data science team, establishing practices and capabilities that would define Vetted's future.

The data science team began tackling increasingly complex challenges. They weren't just ranking products—they were understanding the entire landscape of product reviews, prices, and user preferences.

Tom built automated product discovery to trace references across the internet, expanding what started as a few hand-rolled Python scripts into carefully engineered data lakes and distributed compute pipelines built on Kubernetes, Airflow, and Spark. Sean's team leveraged this data to extract meaningful signals about the ever-changing product landscape.

Infrastructure followed function—as the data science team's ambitions grew, Tom's team scaled out the infrastructure to meet the demand. As Sean's team pushed boundaries, Tom ensured the systems could support their innovations.

Data Science Drives the Critical Pivot

Despite technical progress, users struggled with navigation. They searched for broad categories like "yoga accessories" rather than specific products. While the team explored vector search and embeddings, it was the data science team that identified the transformational opportunity.

During an internal hackathon, Sean and his team created something remarkable: a chatbot that could leverage all their years of data science work—to solve all their discoverability problems—through natural conversation.

This wasn't just adding a chat interface. It was recognizing that large language models could unlock the value of everything they'd built in a fundamentally new way.

The chatbot demo was so compelling it drove a company-wide pivot. Sean's team had shown that data science, properly applied, could solve user problems that pure infrastructure never could. This became the new Vetted.

The Data Science Legacy

When Yahoo acquired Vetted in 2025, they weren't merely buying code—they were buying a data science organization. Vetted's core value proposition were the process and engineering practices that Sean introduced, founded on data science fundamentals of analysis and reproducibility. These processes, crucially, could only have been realized with the right infrastructure to support it. Sean's data science team built Vetted's innovation on data, with raw materials harvested by the machinery of Tom's scalable infrastructure.

The key transformation was putting data science at the center, with infrastructure evolving to support increasingly ambitious AI applications.

Applying Data-First Thinking at Pure Math AI

Now, as CEO of Pure Math AI, Sean brings this data-first perspective to their new venture, with Tom as CTO ensuring the infrastructure can support ambitious AI goals.

Their work at Vetted demonstrated a crucial lesson: successful AI companies aren't built on infrastructure alone, nor on algorithms alone. They're built when data science drives the vision and infrastructure evolves to support it.

Their journey shows that the most valuable AI systems emerge when you:

• Put applied data science at the core of company strategy

• Build teams that can develop, support, and use it

• Create infrastructure that enables rather than constrains AI innovation

• Recognize when new AI capabilities (like LLMs) can transform existing assets

The Yahoo acquisition validated their way of thinking. By making data science central to Vetted's identity—supported by robust infrastructure—Sean and Tom created value that outlasted their tenure. For firms exploring AI implementation, their experience offers a clear lesson: AI transformation succeeds when data science leads and infrastructure follows, not the other way around.

---

Pure Math Editorial is an all-purpose virtual writer we created to document and showcase the various ways we are leveraging generative AI within our organization and with our clients. Designed specifically for case studies, thought leadership articles, white papers, blog content, industry reports, and investor communications, it is prompted to ensure clear, compelling, and structured writing that highlights the impact of AI across different projects and industries. As with any AI-based project, human oversight is employed throughout the content creation process.