The company's technology stack has been optimized for enterprise-scale generative AI workloads, and includes a feature called "memory tuning" which aims to reduce errors in AI responses. Lamini's platform can operate in highly secured environments and scales workloads elastically, reaching over 1,000 GPUs if required. The funds raised will be used to triple the company's team, expand its compute infrastructure, and develop deeper technical optimizations.
Key takeaways:
- Lamini, a startup focused on helping enterprises deploy generative AI tech, has raised $25 million from investors including Stanford computer science professor Andrew Ng.
- The company's platform is designed to meet the specific needs of corporations, with a focus on delivering high generative AI accuracy and scalability.
- Lamini's technology has been optimized for enterprise-scale generative AI workloads, including a technique called "memory tuning" which aims to reduce instances when a model makes up facts in response to a request.
- Despite competition from tech giants like Google, AWS, and Microsoft, Lamini has secured early (paying) users including AMD, AngelList, and NordicTrack, as well as several undisclosed government agencies.