How a Data Processing Challenge at Lyft Inspired the Development of Eventual

While working as software engineers in Lyft’s autonomous vehicle program, Eventual founders Sammy Sidhu and Jay Chia came across a major challenge in data infrastructure that was likely to grow with the advancement of AI.

Self-driving cars produce a considerable amount of unstructured data, including 3D scans, images, text, and audio. Unfortunately, the engineers at Lyft did not have a unified tool that could analyze and process these different data types at once from one platform. As a result, they had to integrate several open-source tools, leading to a prolonged process riddled with reliability issues.

“We had a wealth of talented PhDs and specialists in the autonomous vehicle sector, yet they spent almost 80% of their time managing infrastructure instead of concentrating on their primary applications,” Sidhu, CEO of Eventual, told TechCrunch in a recent discussion. “The majority of their hurdles were centered on data infrastructure.”

Sidhu and Chia played a role in building an internal multimodal data processing tool at Lyft. When Sidhu began seeking new job opportunities, he consistently received questions about developing similar data solutions for various organizations, which led to the idea for Eventual.

Eventual has launched a Python-native open-source data processing engine called Daft, designed to effectively manage multiple data types like text, audio, video, and more. Sidhu stated that the goal for Daft is to innovate unstructured data infrastructure similarly to how SQL transformed tabular data in the past.

Founded in early 2022, just prior to the introduction of ChatGPT and ahead of widespread acknowledgment of this data infrastructure need, Eventual released the first open-source version of Daft in 2022 and is set to introduce an enterprise product in the third quarter.

“After the rise of ChatGPT, we noticed a spike in other developers creating AI applications using various data modalities,” Sidhu remarked. “As a result, the utilization of images, documents, and videos in applications surged, causing a significant increase in demand.”

Although the initial idea for Daft emerged from the autonomous vehicle industry, its applications span across numerous sectors dealing with multimodal data, including robotics, retail technology, and healthcare. Eventual counts Amazon, CloudKitchens, and Together AI as part of its client roster.

Recently, Eventual completed two funding rounds within a span of eight months, starting with a $7.5 million seed round led by CRV, and followed by a $20 million Series A round led by Felicis, with involvement from Microsoft’s M12 and Citi.

The funds from this latest funding round will be used to enhance Eventual’s open-source offerings and support the development of a commercial product that enables clients to create AI applications using the processed data.

Astasia Myers, a general partner at Felicis, shared with TechCrunch that she discovered Eventual through a market analysis focused on data infrastructure capable of supporting an increasing number of multimodal AI models.

She noted that Eventual stood out as a trailblazer in this emerging sector, which is expected to see increased competition, highlighting that the founders’ firsthand experience with data processing challenges was a crucial factor. She also pointed out that Eventual is tackling a growing problem.

As reported by management consulting firm MarketsandMarkets, the multimodal AI sector is projected to expand at a CAGR of 35% from 2023 to 2028.

“In the last two decades, annual data production has increased 1,000 times, with 90% of the world’s data created in the last two years. According to IDC, a significant portion of this data is unstructured,” Myers explained. “Daft is perfectly aligned with this vast macro trend of generative AI encompassing text, images, video, and voice. There’s an urgent requirement for a multimodal-native data processing engine.”