Anna Patterson, a veteran of Silicon Valley, has founded three startups, including the search engines Xift and Cuil, as well as Recall.archive.org, which later became the Internet Archive. She has also held the position of vice president of engineering at Google and founded Gradient Ventures, an AI-focused seed fund. Patterson’s entrepreneurial journey is far from over, as she has recently embarked on a new venture.

According to Patterson, the idea for her new startup struck her a few years ago, but she was uncertain if she had the energy to start another company. However, after being diagnosed with breast cancer in 2023 and taking time off, she realized she had a choice: return to her old life or start anew. It’s clear that she chose the latter, as her new startup, Ceramic.ai, aims to provide foundational AI training infrastructure for enterprises to train large language models more efficiently.

Ceramic.ai claims its model can use long contexts, work with any clusters, and help models scale by 100 times. Patterson’s inspiration for the company stemmed from the realization that existing infrastructure for building large language models was overly complicated and had too many variables, making it difficult for enterprises to adopt at scale.

Patterson noted that when trying to scale infrastructure by 10 times, it usually works without issues, but when attempting to scale it by 100 times or more, it’s necessary to rethink the approach. She asked herself if the current infrastructure was suitable for the next 10 years, which led to the creation of Ceramic AI.

“The thought process behind Ceramic AI was to create a solution that would enable enterprises to train large language models more efficiently,” Patterson explained. “I realized that the existing infrastructure was not designed to handle the scale and complexity of modern AI models, and that’s where Ceramic AI comes in.”

Patterson co-founded Ceramic AI with Tom Costello, the company’s chief scientist, in January 2024. Although the startup has been operating in stealth mode, it has already secured partnerships with notable companies like AWS and Lambda. However, it has not yet generated revenue. Patterson’s strategy is to establish credibility, awareness, and trust with potential customers before focusing on sales.

Ceramic AI recently raised a $12 million seed round led by NEA, with participation from IBM, Samsung Next, and Earthshot Ventures. Patterson chose NEA to lead the round due to the firm’s technical expertise. The funding will primarily be used for sales and continued development.

After founding three startups and holding senior roles, what sets this venture apart for Patterson? Her approach to building startups has changed since her time as a VC investor. She now feels more time-pressured, which she believes is a positive factor, as it encourages her to get the product in front of customers sooner and gather feedback.

Patterson noted that her experience as a VC investor has taught her the importance of moving quickly and being responsive to customer needs. “When I was younger, building companies, I felt a sense of relaxation, just focusing on engineering without sales,” she said. “Now, after being a VC for several years, I feel more time-pressured, which I think is a good thing.”

This sense of urgency is particularly relevant for Ceramic AI, as it faces competition from other companies, such as Together AI, which has raised over $530 million in venture capital, and MosaicML, which was acquired by Databricks for $1.3 billion.

For now, Ceramic AI has its work cut out to establish itself in this rapidly evolving market. With its innovative approach to AI training infrastructure, the company aims to make a significant impact in the industry.