SCM SPECIAL........ GEEKS ON THE GO

SCM ... GEEKS ON THE GO





Scientists from diverse backgrounds, including a neurogeneticist and a particle physicist, are building the next generation of technology for the logistics industry

How can the principle of atoms arranging them selves to form stable metal help pack boxes inside larger containers? Can the pattern of insects walking from their colonies to gather food teach you something about cleaning GPS coordinates for easier route-mapping?
Turns out that the inherent chaos in nature and learnings from particle physics are helping shape new technology for the all-important logistics industry.

Ecommerce-focused logistics firms and online marketplaces are hiring senior scientists previously engaged in top-end research to focus on their core requirement--the efficient delivery of goods. That involves a range of functions, from creating algorithms for identifying where warehouses should be located to geocoding and making last-mile delivery of products smoother.

Pushkar Paranjpe, a neurogeneticist, used to spend hours studying the neural circuitry of walking in fruit flies at the National Centre for Biological Sciences in Bengaluru.He's now a data scientist at logistics technology company Locus, where he designs three-dimensional models for efficient packing of boxes inside larger containers.

“I used to employ genetics to unravel the workings of the brain. What was surprising to me after I came to Locus was that most of the techniques we employ to solve optimization problems derive heavily from natural occurrences around, including biology,“ says Paranjpe.

Suvayu Ali didn't have any exposure to logistics before he joined ecommerce-focused Delhivery as a data scientist in January, and learnt about the industry by sitting through meetings and asking “stupid“ questions. As a PhD student earlier at the Foundation for Fundamental Research on Matter in the Netherlands, he was part of a Large Hadron Collider team researching the subatomic particle called the Higgs Boson, or the God particle. He returned to India when his visa expired and began searching for a job as interesting.

Ali finds that the challenging part of using academic knowledge to predict situations on the ground is that unlike with science, real-life uncertainties cannot be quantified beforehand.

“This Diwali we carried twice the number of monthly volumes. You can't possibly ramp up your facilities for a single month,“ he says. “The project I am working on looks at how we can serve additional orders by mapping the network and removing bottlenecks. It is still under research and we should be ready for the next Diwali.“

Delhivery's data sciences team is developing multiple advanced logistics products, including an address locator, a network planning tool, a product category mapping system for freight shipments, and another for understanding consumer behaviour.

The growth of India's multi-billion dollar ecommerce industry that includes online marketplaces and startups promising quick delivery of groceries and food is dependent almost entirely on warehousing and logistical efficiency. Haphazard townplanning, lack of accurate digital maps and poor road and transport infrastructure pose insurmountable challenges requiring the likes of Paranjpe, Ali and Kabir Rustogi to develop sophisticated workarounds.

Rustogi, a former senior lecturer of statistics and operational research at the University of Greenwich, too, joined Delhivery in January, seeking to solve real-life problems. He soon learnt that what seems great in theory and on screen isn't how things work on the ground.

“One of the products we are working on tells us where to locate our fulfilment centre (warehouses or delivery centres) so that it is close to the demand area. Ideally, Connaught Place (in central Delhi) might be a great location to build one but the real estate costs are (unaffordable). The area manager who has knowledge of the city will black out regions where it is not feasible to build these centres,“ says Rustogi.

Logistics companies have access to huge amounts of data from online marketplaces as well as their delivery centre employees and delivery personnel. Collating these streams of information requires specialised skills in parsing and processing data.

As a computational biologist working on HIV vaccines at The Scripps Research Institute, Shantanu Bhattacharyya was trained to do just that. Working on geocoding at Locus, he says he uses a similar tool set but for a different application.

“It turns out that Google is trying to be as general as possible. When you query Google on an address, the result is the most relevant, not the most accurate. We care about accuracy. Making decisions on which part of the address is important and which is not is very important,“ says Bhattacharyya.

“It is very interesting to know that what is useful to a person is not useful to a computer. Every city and every client and country is different; you cannot have a one rule for all.“

An engineer's perspective on evaluating and solving a problem is as important for a logistics company to solve specific glitches. For Pankaj Batra, vice president (products) at Zomato, the pain points of small restaurant owners who had no visibility on delivery timelines made him think about the last-mile problem. Batra joined Zomato this year when the food aggregation platform bought the logistics technology business of his company Sparse Labs.

“Most delivery boys do not have smartphones. In order to make it easy to track them, we had to outsource hardware from China. After a delivery is completed, a delivery boy can give a missed call to a unique mobile number, which sends a message to the restaurant owner. The owner can track the mileage and location of delivery boys on a simple Android app,“ says Batra, also an avid blogger on personal finance.

Zomato now offers the technology to its partner restaurants, but not yet to its consumers. Rival Swiggy offers real-time tracking for its customers, a gap Zomato is rushing to close.

Even within Ekart, the logistics unit of Flipkart, the architect building an end-to-end logistics network comes from a different team, having built the payments data platform for the online marketplace.

“The product has a business problem to solve,“ says Jagadeesh Huliyar, a senior member of the team developing the logistics network. “The real-time system reduces delivery time and lets you take advantage of economy of planning. For example, if you do not have a full-truck load but have to send shipments from Delhi to Bengaluru, I can plan the route so that the truck makes a stopover at Mumbai delivering additional loads.“

The best people to solve logistics problems need not be from a related space or training, says Saikiran Krishnamurthy, head of Ekart. “The core skills we look for at Ekart are indepth understanding of technology and product, entrepreneurship skills and thinking out of the box,“ he says, adding that the primary value such people bring is the ability to re-imagine the supply chain through technology. That's the primary requirement of logistics companies.

“The people we haven't hired for our team here till now are those with structured 10-plus years of industry experience. The ability to solve problems which deal with uncertainty and flexibility is not something people from corporates are trained to do,“ says Santanu Bhattacharya, senior vice president--data, technology and products, at Delhivery, which he joined last year. Bhattacharya, too, comes from a rarefied background.A Phd from University of Maryland who also worked on a NASA project as part of that, Bhattacharya previously focused on building mobile products at Facebook for emerging markets.

Payal Ganguly and J Vignesh

ET 9DEC16