The explosion of generative AI has created an unprecedented demand for multimodal data processing. Every company building AI applications needs to process massive amounts of text, images, documents, and video - but they're stuck using tools designed for web clicks and bank transactions.

I spent years doing computer vision research and building self-driving car systems at DeepScale (acquired by Tesla Autopilot), and Lyft Level 5. My co-founder Jay Chia built ML data platforms at Freenome and Lyft. We both hit the same wall: processing multimodal data at scale was immensely painful.

We watched brilliant engineers waste months forcing Apache Spark to process images and LiDAR data. Today, we see the same pattern playing out at AI labs training foundation models, at startups building document pipelines, and at enterprises deploying AI applications.

Traditional engines fail because modern AI workloads are different. They run custom models, hit external APIs, and process wildly diverse data types. A 0.1% failure rate that's acceptable in testing becomes catastrophic when processing millions of files in production. 

That frustration Jay and I experienced became conviction. This isn't just workflow inefficiency; it is systematically holding back innovation across the entire industry.

So we built something new: Eventual

Eventual is creating a new foundation for AI infrastructure that’s purpose-built for multimodal data. Our engine handles the inherent messiness of multimodal data as a feature, not a bug - making it possible to query petabytes of images and video with the same simplicity as SQL tables. In fact, our open-source engine Daft is already processing petabytes of multimodal data daily at companies like Amazon, CloudKitchens, Essential AI, and Together AI – these are mission-critical workloads spanning autonomous vehicles, recommendation systems, AI model training, and enterprise data processing.

Eventual’s mission is to build generational technology that makes data processing across all modalities simple, reliable and performant regardless of scale. Today, we’re sharing that we've raised $30M in total funding to support our pursuit of that mission. Our $20M Series A was led by Astasia Myers from Felicis, with strategic participation from M12 Ventures (Microsoft's venture fund) and Citi. Our Seed was led by Brittany Walker from CRV, with support from existing investors Y Combinator, Essence VC, and Array Ventures.

This funding unlocks our ability to hire the best engineers and build the data infrastructure the AI era demands. We’re also launching our early access sign up today.

Our Team

We're currently a team of 18 who've built the infrastructure powering today's largest-scale systems. Our team includes the developers behind Databricks Photon, GitHub Copilot, Pinecone's vector database, Render, and AWS PartiQL; we've built the systems that have processed exabytes.

We're looking for engineers who want to solve hard problems that matter.

Specifically:

• •

  Distributed systems engineers who want to solve problems like: "How do you optimize queries that mix 1KB JSONs with 1GB videos?"

• •

  Product engineers who want to define DevX of working with data in the AI age

• •

  Developer advocates who can show the world a better way to process multimodal data

We're small enough that your work directly impacts our trajectory, funded enough to tackle ambitious technical challenges, and growing fast enough that early engineers will become technical leaders.

Check out our open roles here

Join Early Access

We're opening the waitlist for Eventual Cloud today - the first production-ready platform built from scratch for multimodal AI workloads.

With Eventual Cloud, you can:

• •

  Query petabytes like megabytes - Process millions of documents, images, or videos without infrastructure headaches

• •

  Ship AI features in days, not months - Stop debugging OOMs and start building products

• •

  Write Python that just scales - Your laptop code runs on hundreds of nodes without changes

Built on our open-source Daft engine, Eventual Cloud runs in your cloud with enterprise-grade security and reliability.

The AI era needs data infrastructure built for AI workloads. Not adapted. Not retrofitted. Built for it.

Sign up for early access