7 weeks with Oda: Summer Internship 2021

We are nine interns from three different disciplines who took to the offices of Oda at Nydalen for seven weeks. Over the summer, we have been working on two different cases in two different teams. During this time we’ve been challenged to solve realistic issues that Oda and their users experience every day. Although it’s been challenging, it has given us the opportunity to test our skillsets and learn a great deal, while also working in cross-functional teams. We want to share with you what an internship at Oda is like.

Onboarding 🧒

The first week at Oda was packed with information and onboarding sessions. We had these sessions together with other new employees also experiencing their first week at the Oda. We also got to visit the fulfillment center at Lørenskog where we witnessed the magic firsthand. We got to see how the whole machinery works, and how every small piece is instrumental to the final product: namely, fast and efficient home delivery. After seeing the magic behind the scenes, we joined the drivers on their route delivering groceries for a day. Joining the drivers was especially important for the logistics team since they were going to make a solution for them. It was great fun to see how happy the customers were when we delivered their groceries.

During the first week of onboarding, we also had presentations and conversations with several of the co-founders and other key people in the company about their values, history, workflow, and future plans. 🚀

Flow 🧑‍🏭

During onboarding week, we learned about Oda’s official way of working, called Flow. This method allows Oda to work on key tasks in small teams while allowing easy coordination and cooperation with other teams. The summer teams incorporated this workflow into our summer as shown below (or at least we tried to 😆). One Flow period consists of one week of Flex and two weeks of Focus; this can be described as one iteration. In the Flex period, teams take time to talk to other teams, research, do interviews with end-users, and deciding their goals for the next period. Focus is when all the teams have time to sit down, get things done, and reach their goals.

The interns occupied a corner on the 5th floor in Oda’s office at Nydalen. We had a very flexible work schedule where we coordinated work hours within the teams. This was very nice as it provides the opportunity to work from home/the cabin/the beach when we wanted, and if we needed to come in late or leave early, we’d just tell our team and it’s okay. As mentioned at the beginning, our internship lasted for seven weeks, but we had the opportunity to take a week’s vacation whenever we wanted.

Access to the right people

Since our team consisted solely of interns, we were provided with buddies and mentors. These people were available for questions, syncs, and 1:1 meetings to talk about whatever. As we touched on different technical areas of the organization, we would ask questions in open channels on Slack and were able to get in contact with the right people. For example, we could ask how you get an average of three numbers, and two PhDs would give us a detailed explanation.

Working cross-functionally, it was really important for each discipline to have access to great resources and mentors — especially for the designer and data scientist on each team. They had very frequent and detailed sparring sessions with experts from their respective disciplines. This allowed them to see exactly how Oda works in diverse teams and gave them great help and insight into how to work with stubborn engineers. 👨‍💻

Another great way of getting to know the people at Oda was the weekly “Random Coffees”. Each week we were randomly paired up with another employee from Oda to have a 30-minute chat. This was a cool way of getting in touch with people who worked in different disciplines than ours, and to learn more about other parts of the organization. ☕

The cases

On the very first day of the internship, we were presented with the two cases we were going to work on during the summer. We were also split into two cross-functional teams. The two cases each focused on different parts of the organization: the first focused on the logistics part, more specifically towards delivery and drivers; the second was related to the shop platform, with more focus on the customer experience. Oda put a lot of effort into finding cases that were relevant and could create value for them. 💡

Case 1 — Better Parking

The logistics team consisted of Brage, Markus, Daniel, Andrine, and Peder. Our goal was to make it easier for the drivers to park sensibly in suburban areas. Oda sometimes gets feedback from customers that are unhappy with where Oda trucks park when delivering. Currently, no good ways of relaying information to drivers about where to park exist, apart from writing notices in a text field. In addition, experienced drivers don’t have any way to transfer their knowledge to new drivers, which we also found out in the qualitative interviews our designer facilitated. After gathering information from the drivers and Oda, we started to design and implement our solution in Oda’s codebase.

Our first Flow consisted of facilitating relevant data gathering. We recruited 20 test drivers that agreed to share their positional data when they were 50m away from their delivery point, as well as to test our solution. This meant we could push things to production without affecting all drivers. When the drivers delivered the goods, we made it possible for the drivers to rate the parking in the application the drivers use. This made it possible to measure where good parking spots are, and store that information for later use. To be able to scale our solution we made a profile page for the drivers where they could opt-in if they wanted to contribute with parking ratings. Since GPS-data is notoriously inaccurate, our data scientist made a model that gave us a more exact position.

In our last flow, we started to make the actual product. Drivers are presented with information about the next delivery location in Oda’s already existing app, and we decided that was where our Parking Assistant should live as well. If the driver struggled to find a place to park, we implemented so that he or she could press a button to see a map with good parking spots around their delivery point, the entrance door of the building, and their own current position. This solves a real problem for Oda, and everything we did is something that is almost ready ™️ to be used. 🎉

Case 2 — Your Taste

The shop team consisted of Giske, Wictor, Elizabeth, and Magnus. Our goal was to make a feature in the shop app that could present customers with better product recommendations. In turn, this feature would provide Oda with enhanced insights into customers’ shopping habits and product preferences. We wanted to create a feature that made the customers’ shopping experience quicker, smoother and more fun! 🥳

Our first Flow period consisted of spending a lot of time brainstorming ideas and gathering information. As Oda already had a lot of data on customer insights, as well as great resources like different product recommendation algorithms, we spent time mapping out what resources we could use in our own implementation. We also spent time getting familiar with the different technologies that Oda uses. Let’s just say that the code base of the shop-platform is not particularly small (😅). In parallel, as we continued exploring ideas for our feature, we also started implementing backend features that we knew were necessary.

In the second Flow period we started narrowing down to a concept that we were all happy with. We continued to implement features, and worked towards an MVP. Magnus, the data scientist, started training and testing his ML-models while the developers, Wictor and Giske, focused more on the implementing frontend features. Elizabeth, the designer, tested our concept on relevant users and made adjustments. As the demo-day got closer, there were a few intense work days getting our feature ready, but we made it just in time! Our concept is called “Din Smak”. “Din Smak” is a feature within the shop app where customers are presented with products recommended by our ML-model. They can then choose whether to buy, like, or dislike the presented product, and get updated recommendations the following day (sort of like Tinder, but for food items 😉❤️). The goal with this concept was to help customers to discover items fitting their preferences that are often hidden unless searched for in the current store.

Midway through and also at the end of the internship, we held demonstrations of our projects. It was scary but fun! We’d worked hard to make something that could be of value to Oda. To our relief, the people at Oda seemed to be very impressed with what we had accomplished during our seven weeks. 🙏

Social

At the office is a social area with board games, table tennis and drinks as well as a rooftop terrace, so it was easy to stay after work to have fun together, as well as with others who were there. 🤗

Thank you, Oda, for letting us get to know you ❤️