Super Data Science: ML & AI Podcast with Jon Krohn · Jon Krohn

957: How AI Agents Are Automating Enterprise Data Operations, with Ashwin Rajeeva

·1 hr·10 clips

Data pipelines do not sleep. Jon opens with the pain every data team knows: systems break, errors pile up, and someone gets pulled in at 3 a.m. The idea here is cleaner and a little unnerving. A pipeline could detect an error, rewrite code, and redeploy, then hand the decision back to a human before anything risky happens. Ashwin Rajeeva joins from the Bay Area to explain how Acceldata, where he is co-founder and CTO, is building around enterprise data operations. Jon notes that the company has raised over $100 million, then uses that to frame why data teams are paying attention. The episode keeps coming back to one point: autonomy still needs a gate. Ashwin does not sell agents as something that should run loose and call it progress. The better model is an expert who can approve, reject, review, ask for better work, or request a rewrite. That changes the staffing picture. Instead of engineers repeating the same operational fixes, one expert can direct a larger set of agents. Jon places the platform in large enterprise systems, across petabytes of data. Ashwin grounds it in work Acceldata has done since 2019, including technology shaped by customer conversations and earlier data management problems. Quality checks make it concrete. He talks about validating a million rows every hour, which says a lot about the kind of reliability the system needs. Metadata is useful, but it is not the whole story. Ashwin says the difference is getting to the data underneath, with the ADM platform using a data plane in the customer's environment instead of only pulling cloud-side metadata into a service.

As heard by us

Enterprise data autonomy here means agents, guardrails, and human approval at the point where changes go live.

Ashwin Rajeeva treats enterprise data operations like a control problem: let agents spot errors, rewrite code, and redeploy, but keep a human at the gate.

Read the full review in PlayNext →

Why you'd press play

Want AI agents that can fix data pipelines without waking engineers?

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Super Data Science: ML & AI Podcast with Jon Krohn