jobs in the wild.

Your model is 99% accurate and completely useless. Both are true at the same time. When? If you answered before you finished reading the question — keep going. This is for you. Most AI roles ask whether you know how to use the latest models. I am more interested in whether you can build something that still works when the data is messy, latency matters, regulations apply, and mistakes have real consequences. At DCB Bank, we are building AI and decision systems on our own data and infrastructure—not simply placing an interface over someone else’s intelligence. I am looking for people who can: Train, evaluate and deploy models end to end Read research papers—and challenge their assumptions Think seriously about precision, drift, explainability, latency and production reliability Work comfortably at the intersection of data, technology and banking Take ownership of a problem until it is genuinely solved The problems are not academic. They involve fraud detected or missed, credit decisions made well or poorly, customers understood or ignored, and operating processes that can be fundamentally redesigned through intelligence. This will be a small team with significant ownership. There will be ambiguity, difficult constraints and no shortage of problems worth solving. 📍 Bengaluru · Hyderabad · Mumbai Please do not begin with a generic CV. Send me: The most interesting model, system or product you have built What did not work What you changed after it failed If you are more interested in building AI than merely using it, DM me or mail your CV to shashwat.kumar@dcbbank.com #WeAreHiring #MachineLearning #AppliedAI #MLEngineering #BFSI #Bengaluru #Hyderabad #Mumbai #DeepLearning #AppliedAI #BFSI