About
I am Akash. I live in Dallas and work for the city, where I lead its AI work.
I started out as a chemical engineer at IIT Delhi. The part I actually liked was modelling and simulation, which is the business of writing down what you think a system is doing, running it, and finding out how wrong you were. I did economics next, at the Delhi School of Economics, which is the same activity with people in place of reactors and considerably worse data. Then a masters at Michigan, by which point the tools had been renamed machine learning.
Now I build AI systems for the City of Dallas. One of those is Dear Dallas, a public assistant meant to let residents find city services without first having to learn the org chart, and a number of less glamorous things involving police records, emergency operations, and dashboards that city council reads. Government work has a particular texture to it. There is never enough hardware, the data lives in nine places that disagree with each other, and the one person who knows whether your output is correct has forty other things in their queue that week. I have come round to thinking those constraints are more interesting than they sound, and that most of what gets written about AI assumes none of them apply.
This site is where I work that out in public. Labs is the technical half- experiments, mostly about evaluation and how you know whether a system is any good. The house rule is that the design and the prediction get published before the data is collected, so that being wrong is the point of the exercise rather than an embarrassment in it. The main blog is older, less careful, and about whatever I was reading at the time.
Before all this I spent a few years in behavioural economics, and I have never really left. The question that got me into it is still the one I find myself asking at work- not what the optimal answer is, but what people will actually do when you put it in front of them.
If any of this overlaps with what you are working on, I would like to hear about it. I am slow to reply, but I do reply.
Email · LinkedIn · Twitter · Experiments
Where I have been
- Dallas City of Dallas AI Lead & Senior Data Scientist
- Ann Arbor University of Michigan MS, Information Science. ML and Behavioural Econ
- New Delhi Indian Statistical Institute Data science consultant. Causal Inference.
- New Delhi All India Institute of Medical Sciences Data science fellow. Health expenditure.
- New Delhi Delhi School of Economics MA, Economics. Applied statistics and modelling.
- New Delhi Indian Institute of Technology Delhi BTech, Chemical Engineering. Modelling and simulation.
All views my own. Nothing here represents the City of Dallas.