Quick Thoughts on my "Practical AI for Supply Chain" talks in India

Opex Analytics had a large team in India by early 2020. But, for various reasons, I never went.

I made up for the lost time with a visit this January. I took my son, and we had a mix of seeing friends and colleagues, making new friends, and sightseeing.

Like many others, I’m bullish on India and think the next few decades could belong to India as the previous ones belonged to China. I wanted to visit to get a feel for the country and the people.

So, as part of the trip, I worked with my friend and colleague Sarang Jagdale to arrange talks for data science groups at Chennai Mathematical Institute (CMI) and IIT Bombay.

Our talk at CMI

The talk was “Reinventing the Supply Chain with Practical AI: Nine Trends and Lessons.” We started with traditional supply chain algorithms (and I focused on network design as an example). We then covered the different definitions of AI and how new algorithms can solve new problems and complement traditional ones.

I was nervous. Most of the material had been refined for a US audience. So I wasn’t sure it would play in India. My nervousness was misplaced. The talks played well, and I enjoyed the time with the students and professors.

Here are my quick thoughts on the talks.

The more I talk to data science groups, the more I’m reminded that it is important that we introduce them to optimization. Optimization plus machine (and deep and reinforcement) learning are perfect for solving business problems. Unfortunately, optimization (and some traditional OR algorithms) tend to be overlooked in the data science community. INFORMS is trying to address this with Laura Albert’s push to get closer to the larger AI community.

I was worried that my network design examples were too US-focused. However, the fear faded when I learned that at least one student was going to work in the network design group of a large US firm! Many more were going to US firms. So, the examples felt fine. I’m hoping that, in time, I’ll have many more India-based examples.

Sarang had a good take on IP (Intellectual Property). He thought this audience would want to understand IP in the supply chain domain. IP isn’t just the algorithms; many of those have been around for decades. Instead, IP is the combination of algorithms + engineering + usability (good design) + deployability (to get it into the hands of users).

I have a lot to learn from India. The audience’s questions on my network design material made me want to clarify some points and build new cases for India. In addition, I have a lot to learn about same-day delivery. When sightseeing in Mumbai, our guide took us to the meeting point of the dabbawalas. This is Mumbai’s ~100-year-old organization for getting homemade lunches from homes to schools and offices throughout one of the world’s largest cities. It is a case study in excellent logistics written up by HBR. And, from what I now understand, there are many other great examples of same-day delivery.

This dallawala saw my interest in the operation and insisted we take a picture.

A dallawala bike loaded for delivery.

I’m looking forward to my next trip and doing more talks.

---

If you want to subscribe for free and get these posts in your email, click the button below.