Scale AI’s former CTO launches AI agent that could solve big data’s biggest problem

Isotopes, a new startup, has officially launched with a promising $20 million seed funding aimed at tackling a long-standing issue in data analytics. For years, the challenge has persisted: the experts who manage big data systems are often not the same individuals who need to analyze that data. With the advent of large language models (LLMs), business managers can now pose questions to their data using natural language, bridging that gap. The company introduces Aidnn, its innovative AI agent designed to sift through various data sources, including finance applications, ERP systems, CRM platforms, and cloud storage, to generate insightful responses and draft detailed planning documents. Despite the existence of numerous business analytics tools, Isotopes stands out due to the impressive backgrounds of its co-founders, leading to a product that is both sophisticated and effective. CEO Arun Murthy shared that the startup has already filed for ten patents, showcasing their commitment to innovation. Murthy's experience dates back over two decades when he was involved with Yahoo's Hadoop project, which ignited the Big Data movement in the 2010s. After co-founding Hortonworks, which later went public, he witnessed firsthand the challenges posed by evolving cloud technologies that eventually led to its merger with Cloudera. During his tenure at Cloudera, Murthy encountered persistent data access issues, recalling moments when executives struggled to provide answers to analysts due to a lack of data availability. In 2021, after leaving Cloudera, Murthy connected with Alexandr Wang of Scale AI, where he served as CTO, gaining invaluable insights into model development. However, a conversation with fellow Hortonworks alum Prasanth Jayachandra inspired him to embark on a new venture alongside Jayachandra and Gopal Vijayaraghavan, another former colleague. Together, they established Isotopes in late 2024, with funding led by Vab Goel of NTTVC. The capabilities of Aidnn extend beyond simple queries; it can locate data across various platforms and clean it for analysis. This AI agent retains contextual memory, which aids in executing complex tasks. For instance, when tasked with creating a report on monthly recurring revenue trends, Aidnn can undertake a multi-step process that involves data extraction, cleaning, normalization, and aggregation, while transparently showing its reasoning and highlighting any data anomalies. Moreover, Isotopes assures enterprises that they can utilize the agent without sharing sensitive data with the AI model developers. Nonetheless, the startup is entering a competitive landscape, facing established players like Salesforce’s Tableau, which are also advancing their own AI solutions. As Isotopes continues to innovate, it aims to redefine how businesses interact with and leverage their data.