There is no general consensus as to what a data scientist is or how to become one. Is she a statistician, an operations researcher or a broker between data analysts and developers? Perhaps identifying a data scientist’s work by her common tasks is not the best approach. Instead, why not run through a day on the data science team here at Gousto? This post is about a recent trip the team made to the company warehouse in Spalding.
Squinty-eyed and coffees-in-hand we met at the station and made our way to the 7:30 train to Peterborough, from where we were set to take the next train on to Spalding. For those who haven’t recently ventured out to Lincolnshire, this is what the local public transport looks like:
Walking into the Spalding office, we found the rest of the team already fully in work-mode – their day typically starts at 8AM. No time for a casual chat at the coffee machine.
At the start of every day, each member of the team typically takes an hour to get a head start on the day and work on BAU (Business As Usual tasks). This means checking the results produced by our algorithms driving or supporting for instance forecasting, inventory management or marketing. Timely review of each other’s work on GitHub is an important part of our team’s workflow. With the impending Stand-up on our minds, we often try to review any outstanding pull requests so we can all push ahead and close tickets to maintain a solid burn-down like in the chart below.
The Tech team at Gousto work using the Agile methodology, and so do we in data science. One of the so-called “ceremonies” in Agile is the daily stand-up. Stand-up is a quick 15-minute meeting to discuss the previous day’s highlights and what we’re planning to work on for the day.
We’re constantly innovating the process that starts the moment a customer orders their weekly recipes and ends the moment the recipe box is delivered to their doorstep. Naturally, the warehouse is crucial to making this process run smoothly. New ideas are frequently implemented in Spalding, meaning this time round there were new things for us to see in action, as always. What’s more is that walking around and seeing how a box goes from empty to filled, stickered and ready to ship is a great moment to discuss and imagine what’s next? We also got stuck in and did some actual picking. We hope we didn’t make any picking errors! Disclaimer: we did not forget the Health & Safety standards (photo proof below).
Unlike our office in London, the Spalding warehouse is a place where ingredients are packaged and processed for delivery, not cooked and prepared for a midday feast. Packed lunch it was. Fortunately we could sit outside in the sun.
Fueled by leftover Easter chocolates from the Gousto Marketplace we did a deep-dive into recent work on the suite of secret-sauce algorithms that optimise our warehouse operation. Armed with markers we did our magic on the white board. Other colleagues sitting in the same office space joined in on the discussion and contributed their thoughts on implementing the next round of improvements. Our discussion briefly derailed as we touched on the physics of jet engines, but we managed to get back on track and talk algorithms, dimensional reduction and feasible solutions.
Being off-site presented us with a chance to think beyond the next few months and discuss longer term plans in line with the company strategy. In case of a data-driven business like Gousto, data science can be instrumental in scaling the business and enabling new data-driven product and service features. For the next 60 minutes, no idea was too outlandish while we pictured Gousto in 2018 and beyond.
After a day of thinking, imagining and planning it was time to head back down to London. We decided to take advantage of the well-known Ballmer Peak and enjoy a beverage on the train to stimulate our brains for some end-of-day creative work.
Rolling into London, our sights were already set on the following day in our familiar office environment, ready to plug away at a new challenge.