Finding the road best travelled

 

Meet Laszlo and hear about the vital work he and his team do to keep our huge fleet of trucks on track, the mysteries they get to be part of solving every day, and the fun they have doing it.

 

“In my experience, 21 years in this field, it’s rare to find a team that’s so close-knit. Helping each other comes naturally and we all know we’re stronger together. The human side is probably what I like the most about my job. There are no egos and I think that’s very precious.

I’m a Software Development Engineer (SDE) in the Hive team. We plan UK-wide routes for several thousand trucks moving goods in bulk between depots and stores every day. It’s important work and so much goes into it, from designing the systems to quality assurance. And there’s no room for error. Our team is made up of a couple of junior and senior SDEs, Mathematicians, and a Cartographer. But even though we’re small, I’m never stressed, and we actually have a lot of fun.

At the heart of the routing algorithms, we’ll always find this ‘travelling salesman problem’ which is basically an optimisation problem. It’s about finding the shortest, cheapest, most sustainable routes between Tesco locations (depots and stores). And on top of that, Tesco has 26 different kinds of trucks, so we’re using different maps for every single one. We’ve got a lot to consider, like if a particular truck will fit under a bridge or be able to turn at a roundabout without getting stuck. It’s easy to solve when you’ve only got ten or twenty cities to think about. More than that, and the problems escalate; we have 2000 locations between which we must find the most optimal routes.

 

Here at Tesco, we’re trying to solve a more complicated version of the ‘travelling salesman problem’.

We have to think about the time-based element, reducing CO2 emissions, cost. If we can optimise something by just 1%, we’re saving a huge amount of money that can be put to better use than burning fuel. So we’re doing good for the company, but for the planet too. And optimisation problems don’t have precise solutions, it’s usually an approximation of what the perfect solution could be. Working on these solutions is very interesting and truly exciting because you could make a discovery any day. It’s one of the best feelings to find a connection between two things or being able to solve a problem in an abstract way.

My team’s impact is big. We’ve seen the statistics, and we’ve got pretty good results. Obviously, we didn’t start the revolution – there have been teams before us, but we can always improve on things. Last year we rewrote a component which was using a large machine that took 80 minutes to crunch some data. We were able to optimise it so that it now takes just three minutes. It’s very satisfying when you can reduce costs and create greener ways of working.

 

This year, we fixed an even bigger problem. We were having to use up to 50 very expensive machines to do the biggest part of our computations. It was taking 12 hours for every run, and we managed to reduce it to 30 minutes, which was a great feeling. The prod-deployment, in which we rewired multiple channels of communication, actually took us 11 hours. It was the longest Teams meeting I’ve ever had. We came up against a lot of issues, but we were laughing all day. I think that’s pretty cool and definitely surprising because generally, when work is hard it can be frustrating, but here it’s fun. And that’s because the guys who I work with are brilliant.

My role has given me so many opportunities to solve mysteries, which is great because I’m curious about everything. We’re currently working on a long-term challenge, an enigma we need to crack. When a graph is configured in a certain way, our computations may slow down dramatically, or even get stuck in infinite loops and we’re trying to understand why. So we’re working as a team to conduct various experiments and we have some ideas and workarounds that can help us with short-term issues, but we want to fix it for good. It’s not a problem I’ve faced before, but it’s not stressful or tense. In fact, these are the kinds of challenges that are exciting for us. So, if you’re a curious person, come and be curious with us, we’ve got interesting mysteries for you here.”

So why not explore our latest Tech opportunities? Click here