Your email is on its way!
To track your gifted subscription, go to your Account Settings 
We’ve emailed your gift link. Want us to send some more?
Yes, Send more
0 of 0 Subscription available
Not now, thanksTo track your gifted subscription, go to your Account Settings
We’ve emailed your gift link. Want us to send some more?
0 of 0 Subscription available
Not now, thanksMrunmayee covers AI and Urban development. She has a background in engineering and journalism and is interested in the intersection of policy with industry.
By continuing, you agree to our Terms of Service and Privacy Policy
“There's no greater gift than thinking that you had some impact on the world, for the better.”
– Gloria Steinem
Our anti-piracy system has flagged your account for suspicious activity and has temporarily paused your account. This may happen due to a number of reasons.
If you think that this was done in error, please get in touch with us at [email protected].
Top Comments by Mrunmayee Kulkarni
14,000 stalled e-buses. How many agencies does it take to fix them?
Fair question. So while the payment is linked to kilometers run, it's not a straightforward 'Pay Rs x for y km run this month.' Under most GCC contracts, there is an assumed minimum availability - typically around 85% of the contracted kilometers. The OEM prices the bid on the expectation that the bus will run roughly that much. If a bus runs less than that, the STU can deduct money for the kilometers not run. But. The amount of penalty an STU can impose is capped. Beyond a limit, additional downtime does not lead to additional financial loss for the operator.
Mrunmayee Kulkarni The Ken, Staff Writer
What does it take to build a new tech city? Ask Karnataka’s neighbours
Furnished office setups around Koramangala or along the ORR do rent out for an average of Rs 92 per square foot
Mrunmayee Kulkarni The Ken, Staff Writer
How AI turned banks’ risk data into advertising
Fair point on hype. But just to be clear, this piece isn't conflating ML algorithms with "AI." This is more about learning systems in BFSI martech. The look-alike model (to target new customers) and uplift model (to predict how likely someone is to respond to an ad or a nudge), for example, are definitely built using ML. But when they are running in a live martech system - constantly taking feedback, tweaking campaigns, updating probabilities - that's AI in action. Without AI, you would have a dashboard. With AI, there's still a human in the loop (for now), but you have a decider.
Mrunmayee Kulkarni The Ken, Staff Writer
How AI turned banks’ risk data into advertising
The AI angle here isn’t about creativity — it’s about velocity and precision. It didn’t reinvent the banks' ad-targeting; it just made it faster, sharper, and endlessly scalable, using the same big data banks have been cleaning up.
Mrunmayee Kulkarni The Ken, Staff Writer
How AI turned banks’ risk data into advertising
Fair. Data enrichment and programmatic targeting are definitely the machinery here. The point though is how AI is amplifying - and will eventually automate - these processes. What once took ages, can now happen in minutes. What used to be a manual mapping of “affinities and behaviour” is now a self-learning loop that discovers, scales, and sells - all in the same breath.
Mrunmayee Kulkarni The Ken, Staff Writer