Former Scale AI CTO Launches AI Agent Startup to Tackle Big Data’s Access Problem

Isotopes, a new AI startup emerging from stealth, has raised $20 million in seed funding and introduced Aidnn, an advanced AI agent designed to tackle one of the most persistent challenges in big data: the gap between data infrastructure experts and the business users who need to make decisions based on that data. Unlike traditional analytics tools that require technical expertise to query data, Aidnn enables non-technical users—such as business managers and executives—to ask complex questions in natural language and receive accurate, actionable insights. The agent can pull data from diverse sources including finance systems, ERPs, CRMs, and cloud storage platforms, then automatically clean, normalize, and analyze it to generate reports, forecasts, and strategic planning documents. What sets Isotopes apart is the deep technical pedigree of its co-founders. CEO Arun Murthy played a pivotal role in the early days of big data. In his mid-20s, he helped build Hadoop at Yahoo, the open-source framework that launched the big data revolution. He later co-founded Hortonworks, which became a public company before merging with Cloudera in 2013. Despite his success, Murthy remained frustrated by a recurring problem: even at a data company, executives couldn’t access or interpret their own data during investor calls, leading to embarrassing moments. After stints at Cloudera and consulting with Scale AI’s Alexandr Wang, Murthy joined Scale as CTO. He described the experience as “getting a PhD at Scale,” where he gained deep insight into how large language models are trained and optimized. In late 2024, Murthy reunited with two former Hortonworks colleagues—Prasanth Jayachandra and Gopal Vijayaraghavan—to launch Isotopes. Their combined experience allowed them to build an agent that doesn’t just retrieve data, but understands context, performs multi-step reasoning, and tracks its own logic. For example, to analyze monthly recurring revenue trends, Aidnn automatically extracts metadata, cleans raw data, joins datasets, prorates revenue, and aggregates results—all while documenting each step, flagging anomalies, and suggesting next actions. Isotopes emphasizes security and privacy. The agent operates on-premises or in private clouds, ensuring enterprise data never leaves the customer’s environment or gets shared with external AI providers. Despite its technical edge, Isotopes faces strong competition. Salesforce’s Tableau and other analytics platforms are rapidly integrating AI agents, while startups like WisdomAI are also targeting the same space. But with its founders’ proven track record in building foundational data infrastructure, Isotopes aims to deliver not just a tool, but a new way for organizations to access and act on their data—without needing a data scientist in the room.