AI, particularly text-generating AI, such as large language models akin to ChatGPT, has gained significant attention. In a recent survey encompassing approximately 1,000 enterprise organizations, 67.2% expressed that adopting large language models (LLMs) is a primary objective by early 2024.
Despite this enthusiasm, obstacles impede progress. The survey indicates that a lack of customization and flexibility, coupled with the inability to preserve company knowledge and intellectual property, hinders many businesses from integrating LLMs into production.
In response to this challenge, Varun Vummadi and Esha Manideep Dinne founded Giga ML, a startup focused on developing a platform that enables companies to deploy LLMs on-premise. This approach aims to reduce costs and safeguard privacy.
Vummadi highlighted the significant challenges faced by enterprises in terms of data privacy and customizing LLMs. Giga ML seeks to address both challenges by offering its own set of LLMs, the “X1 series,” designed for tasks like code generation and answering customer queries. Built on Meta’s Llama 2, these models claim to outperform popular LLMs on specific benchmarks, including the MT-Bench test set for dialogs.
While Giga ML’s models may excel in certain aspects, the focus appears to be on providing tools for businesses to locally fine-tune LLMs without relying on external resources and platforms. The objective is to simplify the training, fine-tuning, and deployment of LLMs through an easy-to-use API.
Vummadi emphasized the privacy advantages of running models offline, a factor likely to resonate with businesses. Giga ML’s mission revolves around assisting enterprises in securely and efficiently deploying LLMs on their own on-premises infrastructure or virtual private cloud.
Acknowledging concerns around sharing sensitive data with vendors, Giga ML aims to provide secure on-premise deployment, customizable models tailored to specific use cases, and fast inference, ensuring data compliance and maximum efficiency. This privacy-centric approach is particularly valued by IT managers at the C-suite level.
Having secured approximately $3.74 million in venture capital funding from investors like Nexus Venture Partners, Y Combinator, Liquid 2 Ventures, and others, Giga ML plans to expand its team and intensify product research and development. The startup’s current customer base includes unnamed enterprise companies in the finance and healthcare sectors.
