Sphinx Emerges with Copilot for Data Science

Sphinx emerged from stealth today with an AI copilot designed specifically for data scientists and $9.5 million in seed money. While the typical copilot targets software developers, Sphinx aims to bring AI technology to bear on common data science tasks within Jupyter and other data science notebooks.

Sphinx is a VS Code extension that leverages a variety of GPT-4 models from OpenAI to help data professionals with a range of data-oriented tasks, such as refining data, verifying AI models, and grounding responses. The software runs entirely on the user’s system and “learns” as the customer uses it.

Rohan Kodialam co-founded the company with CTO Jamie Bloxham after concluding that the latest AI copilots were helping traditional software developers but leaving data scientists and other data professionals asking for more.

“AI is driving a paradigm shift for natural language and code, but traditional data has been left behind,” Kodialam, who is Sphinx’s CEO, stated in a press release. “Our researchers and engineers are aggressively innovating on the interface between AI and data to drive tangible value for our partners across industries including CPG, retail, and financial services.”

Sphinx claims a top spot on the DABstep data agent benchmark for multi-step reasoning

The Sphinx copilot uses a variety of AI and ML models, including the latest reasoning models, to help data scientists with their data science tasks. For example, it can call a classic clustering algorithm to run a statistical analysis, and then use a large language model (LLM) to generate an explanation of how the results were attained. It can leverage a combination of image models and statistical models to spot problems in a graph, and then use GenAI to explain where the code is broken.

Lightspeed led the $9.5-million seed round for the New York company, which included participation from Bessemer Venture Partners, Box Group, K5, Impatient VC, and others. Other backers include Steve Cohen, Naveen Rao, and leaders from Databricks, Windsurf, and Together AI, the company says.

Bucky Moore, a partner at Lightspeed, said Sphinx brings frontier AI capabilities to data analysis and redefines how AI reasons with data.

“Starting with the core workflows of data teams, Sphinx’s agents will continue to handle more of the tedious work that goes into deriving insights from data,” Moore stated. “It’s more critical than ever for enterprises to glean key information from their data to fuel business decisions and Sphinx enables this at record speed. Rohan and Jamie are on a mission to define how enterprises leverage AI to become truly data-driven.”

Sphinx Co-founder and CEO Rohan Kodialam (right) and Co-founder and CTO Jamie Bloxham

Kodialam, who has a Master of Engineering degree in AI from MIT, spent time at the Princeton Plasma Physics Laboratory and Nokia Bell Labs before helping to develop medical devices at Affera Citadel. Kodialam spent seven years at Citadel, where he was head of AI development of the data strategies group before co-founding Sphinix.

Bloxham got a BS degree in computer science from MIT and went on to work at Quora and ScaleAI before joining Rao’s startup, MosaicML as a software engineer. He spent the last two years at Databricks (which bought MosaicML in 2023) before joining up with Kodialam on Sphinx.

The Sphinx Co-pilot is available now. Customers can start out with the free Scout version, which gives them 50 agent units per month and community based support. For more agents and support, customers can move up to the Explorer package, which delivers 200 agent units per month for $20 per month, or select an enterprise package.

For more information, check out www.sphinx.ai.

Pilot vs. Co-pilot: How Startups Will Reshape the Future of Work with AI

Data Prep Still Dominates Data Scientists’ Time, Survey Finds

The post Sphinx Emerges with Copilot for Data Science appeared first on BigDATAwire.