[ad_1]
A new company emerges from stealth today with support from Google’s AI-focused venture fund to help companies package open source AI infrastructure and reduce engineering overhead.
cake Integrates and secures more than 100 enterprise components, including data source adapters (such as Apache Hadoop), data ingestion (such as Apache Kafka), data labeling (such as Label Studio), vector and graph databases (such as Milvus or Neo4j), and generative AI APIs and related tools (such as Anthropy), among many other categories.
This speaks to why Cake has its name – it takes the different “layers” that make up the AI stack, and combines them into a more digestible, production-ready, business-friendly format.
“Big picture problem”
Founded in New York in 2022 by Misha Hersko (CEO) W Skyler Thomas (CTO) – Pictured above – Cake launched last year and is already working with clients like Bioscience AI startup Altis Labs and data intelligence insurtech ping. However, the company has not made much noise in public yet.
On top of today’s official reveal, Kik said it has raised $13 million since its inception. This includes $3 million in seed funding during its formative two years, and a recent $10 million seed round led by Google’s Gradient Ventures.
“We weren’t very secretive; we just built and worked with customers,” Herscu explained to TechCrunch in an interview last week.
Previously, Hersco founded an AI company called McCoy Medical Technologies that focused on machine learning infrastructure for radiology, and He sold it in 2017 For an IT vendor TeraRecon. He later joined Primary Venture Partners in New York as an Operator in Residence, where he pursued his next project by chatting with hundreds of data science and AI executives.
“I’ve made more than 200 discovery calls to customers, asking them what their biggest pain points and bottlenecks are,” Hersko said. “The biggest problem wasn’t just one part of the stack, like setting up a vector database or a data pipeline. There were a lot of different components across a very rich ecosystem. How do you reliably integrate everything and make it production-ready?”
This is what Hersko refers to as the “big picture problem,” and is where his new company enters the fray.
Cake is about understanding the myriad of open source components that make up the modern AI stack, and providing aggregated, managed, open source AI infrastructure for small teams. It’s not about building a business around a single open source project as countless companies have done; Instead, it’s about bundling and serving a selection of open source projects across the entire stack and making them work seamlessly.
Let’s say a large financial services company has millions of documents containing complex financial data, and wants to perform RAG (Retrieval Augmented Generation) against these files to improve the quality of responses to natural language queries. If the off-the-shelf product is not up to the task, or is not suitable for compliance reasons, the company will have to build its own system by installing and integrating multiple different components. This is a time-consuming endeavor that Cake can take care of.
Elsewhere, a hospital may need to create a secure system for analyzing CT scans, or an e-commerce company may want to upgrade its recommendation engine. These are all potential use cases for cake.
“We run the gamut, but I would say our best work is definitely when companies go beyond what you can do with a simple, off-the-shelf product,” Hersko said.
Parallel development
Thomas, CTO, previously worked at IBM as a principal engineer, and most recently was a distinguished engineer and director of strategy at Hewlett Packard Enterprise, which acquired a previous company he worked for called Mab ar.
Thomas says he’s worked across hundreds of projects over the years, with clients large and small, and he noticed a trend permeating almost all of them — that every one of them was using open source tools in some way, and most of them were new to research. Laboratories. However, its use in the enterprise was not easy.
“It takes a long time for even the largest companies to take what comes out of the labs and integrate it into what they do,” Thomas told TechCrunch. “A lot of that is because most of it is not enterprise ready – it may not have authentication and licensing, and organizations have to do it themselves.”
There are similarities to what Keck is going for here. In Europe, we have Finnish proverbs Evina $2 billion unicorn, that does something similar but with a focus on data infrastructure. Perhaps the most obvious comparison is to Red Hat, which IBM acquired for $34 billion, and which is best known for its enterprise-level Linux (RHEL) operating system.
“In the early days of Linux, there were thousands of open source packages that everyone wanted to use, but they weren’t integrated and weren’t secure,” Thomas said. “There was no support model for it, so the red hats of the world made Linux safe for enterprises. We want to do something similar for AI today.”
While there are plans to eventually offer a hosted version of Cake, companies currently have to run it in their own environments. For many, this won’t be a problem because data privacy terms mean they can’t send data outside their own systems anyway. But the hosted version may be attractive to organizations with fewer compliance obligations.
“It would be easier for us if we could control the cloud,” Herscu added.
Aside from lead investor Gradient, Cake’s seed round saw participation from its primary investors Primary Venture Partners, as well as Alumni Ventures, Friends & Family Capital, Corelation Ventures, and Firestreak Ventures.
The as-yet-unannounced $10 million seed round, which closed in April, hints not only at the founders’ backgrounds but also at the company’s appeal. The company is already eyeing its next funding round, with tentative plans to raise money again around mid-2025, Hersko said.
“From a traction standpoint, we look more like a Series A company already. We were able to get there very quickly,” Hersko said. “When we go to Series A, it will probably look more like Series B.”
[ad_2]