Profile

My job links the results achieved by universities to a commercial product used by everyone. Specifically, the purpose of my job is to study the results and scientific publications produced by universities and select what is appropriate for the need of my employer. This means that I have to study new stuff every day, perform some preliminary tests with that, discuss the results with other people and managers and, if the results look good, spend a lot of time developing a prototype. The prototype will be inspired by the original scientific research but adapted to the needs and resources of the employer. Once I finally figured out how the prototype can work and perform a bit better that what we are already using, I give all my software, the instructions and documentation to a team of software engineers. Their job will be to clean things up and make the software easy to use, fast and robust. It is not worth to make prototypes usable by non-technical people when they are not finished, because their structure changes continuously as it is difficult to find out how to make them to work. This is how my job is when working on an “internal project”.

In my specific case things are a little different than the above, as I am employed on a European Union project (EU H2020 SecondHands). In such projects, many universities (from the EU and UK) are working together to perform research on a specific topic, and more and more often companies are also involved to contribute or test the results. In such case, the topic of the project and length are agreed from the start and little changes are happening in due course (while an internal company project can evolve freely). EU projects are more theoretical than internal projects as they are pioneering a new things and concrete applications might not be as direct as for internal projects. Once the project is finished, then I might work on a more “traditional” internal project or on a new EU project.

The EU project I work on focuses on developing a whole humanoid robot which has to help a technician when repairing the facilities of Ocado’s automated warehouses. This has to be a complete system: mechanics, electronics, software, sensors and intelligence, and must be able to safely work with a human. The robot must understand what the technician is doing and anticipate him in providing him help by e.g. collecting a tool from the table and handing it over, asking questions if necessary. This project is a collaborative effort of 4 EU universities plus Ocado, who has to “wire up” all the software elements of the robot and test it in the actual warehouse. A lot of my current work in this project is about messing around on Linux, C++ and Python to make sure that every part of the robot’s intelligence is able to communicate with the rest and to provide meaningful information for the whole system to work.