Why I joined Eventual

For most of my career, I’ve worked at companies building and commercializing open-source Apache projects. That work put me deep into distributed systems and infrastructure, where design decisions directly impact scalability, reliability, and cost. It also shaped how I think about building systems that are broadly usable, maintainable, and grounded in real-world constraints.

Over time, my focus expanded beyond core systems to building managed services for hundreds of customers across a wide range of deployment models and cloud environments. Supporting that level of scale and diversity exposes the limits of one-size-fits-all solutions and forces you to think carefully about abstractions, operational simplicity, and how software behaves in real production settings, without requiring every user to become an expert in the underlying systems.

More recently, I became increasingly interested in how AI and machine learning workloads are reshaping the problems that distributed systems need to solve. We are still in a relatively early phase of a major technological shift driven by AI, and the core infrastructure patterns are far from settled. Multimodal data, large batch inference jobs, and specialized compute place very different demands on systems than traditional data processing pipelines. For me, this felt like a natural next step, an opportunity to work at the intersection of data processing, AI, and infrastructure while the foundations are still being defined.

That’s what drew me to Eventual.

Eventual is building infrastructure to support large-scale AI and multimodal data workloads, grounded in Daft, an open-source Python framework for processing data across modalities. What stood out immediately was the focus on unifying data processing, inference, and AI infrastructure into a single, coherent system. With just a few lines of code, engineers can efficiently process large volumes of data and run inference without having to stitch together multiple specialized tools or manage complex distributed systems themselves.

Daft, in particular, feels like generational technology for multimodal data processing. It brings open source principles to a space that has historically been fragmented and difficult to operate at scale. Rather than treating inference as an afterthought layered on top of data systems, Daft is designing for data processing and inference together, recognizing that this is where many real-world AI workloads struggle today.

From a technical perspective, this was deeply compelling. Scaling distributed systems for batch inference introduces new performance bottlenecks, new failure modes, and tighter coupling between data and compute, especially in pipelines that process massive datasets and then run inference across specialized hardware. Solving these problems well requires systems that are both powerful and approachable. Daft’s focus on simplifying both large-scale data processing and inference addresses a gap that has long existed, and it enables companies to make meaningful use of their data assets without overwhelming operational complexity.

The team and founders were another major reason I joined. In my conversations, it was clear they bring deep experience across both distributed systems and machine learning, particularly around the realities of running AI workloads at scale. Just as important was the opportunity to learn from and collaborate with a highly talented group of engineers who care deeply about the craft. There is a shared emphasis on strong system design, attention to detail, and building infrastructure that holds up over time.

Ultimately, joining Eventual felt like a rare opportunity to contribute to open source software while helping define the next generation of AI infrastructure. I’m excited to work on hard distributed systems problems, to learn alongside an exceptional team, and to help build tools that make large-scale data processing and AI inference simpler, more accessible, and more robust. If you’re excited by hard systems problems, care deeply about open source and craft, and want to help define the foundations of AI infrastructure, Eventual is a special place to do that work.