An interview series that highlights the human side of interesting people in tech
Amina Souihi interviewed by the FT.
What’s your job title?
Ph.D. student in Analytical Chemistry at Department of Materials and Environmental Chemistry, Stockholm University. My project is about developing semi-quantitative methods for non-targeted liquid chromatography coupled to high resolution mass spectrometry (LC/HRMS) using machine learning tools, programming, and multivariate data analysis.
The aim is to analyse samples without any prior knowledge. We use different machine learning algorithms to develop methods that can determine the presence and amount of the compounds existing in the samples. These methods consume less time and money, and they reduce the need for standard chemicals, so it’s more environmentally friendly.
Do you focus on a specific field, like food?
No. I’m developing predictive methods that can be used in different applications e.g. water and food analysis, environmental applications, etc. At the end of my research project, we aim to test our predictive models on water samples.
How do you verify that your machine learning is giving you reasonable results?
The general procedure consists of dividing the dataset into a training set and a test set. The training set is used to train the model, and the test set is used to evaluate the prediction ability of the model. In addition, we use an external dataset to validate the prediction ability of our model. Currently, I’m trying to predict the retention time of the compounds in different chromatographic conditions, which could be beneficial to identify the compounds.
Where do you get the data samples from?
I work in collaboration with master students, and I combine their datasets to test different machine learning algorithms later. And our research group has collaborations with different universities and research groups, so we try to get data from them as well. The data is usually checked from a chemical point of view before starting the data analysis.
Do you check the data manually?
Yes, we have to, sometimes. I use the multivariate models and the predictions to see if a manual check is needed. For example, I double-check the outliers manually — it’s much easier than checking the whole dataset. The analyst needs a chemistry background. Someone with just the machine learning skills wouldn’t be able to see the chemical side.
What do you do on a normal day?
It differs a lot. As a Ph.D. student, I have to take courses, teach as a lab assistant, work on my research project and supervise students. So it depends. Currently, I’m working on my research project; I do all the data analysis. And at the same time, I’m teaching as a lab assistant in a bachelor level course called “Basic analytical chemistry”, doing the labs for the students. [Because of the pandemic] we pre-record the labs and then we watch those videos together with the students, try to answer their questions, and explain how they need to prepare the reports. But sometimes I’m taking courses 100% of the time, which means attending the lectures and work on all the projects given in that course.
That’s a lot 😳
Yes, but when I’m programming in R it doesn’t feel like work or studies — it’s a hobby for me. I don’t count the hours. I can stay up until midnight and then I realise that I’m still working on that code and that I need to go and sleep.
What was the first piece of code you ever wrote?
Oh My God 🤦♀️ That was sooo basic. I took one programming course as part of my undergraduate studies. In the beginning I didn’t like it, to be honest — I felt completely lost. I think I talked to my brother, and he said: “Oh, if you can learn programming that’s perfect, it will open a lot of opportunities for you. It’s something you should learn”. He motivated me a lot. And then I went back home, and I still remember just repeating the exercise that the teacher showed us in the class. I think the first line of code was “print something” and I was extremely happy to have it printed. I can’t remember exactly the exercises, but that was the first thing I ever wrote. In python, at that time.
What was the last piece of code you wrote?
That was to combine data from different labs and to develop different projection models. This direct mapping of data can be used to quantify chemicals for new labs. In other words, we combine the data in one database (into the same scale), and use this information to predict for new labs/instrumentation.
What advice would you give to someone just starting out?
First of all, enjoy your student life! Most of the time we forget to enjoy it, but after you finish your masters you realise that it was the best part of your life. Try to make friends, enjoy your fields.
And the second thing is — even for a chemist or someone in engineering: learn programming languages, because it gives a lot of opportunities. If I didn’t learn python and R, I might not be here, might not have been a Ph.D. student, so it was a big plus for me, and I believe it’ll be the same for many people.
Also: work hard! Try to develop new skills. It doesn’t have to be part of your (bachelor) program. Just be curious, don’t wait for someone to show you or teach you, just teach yourself. Try to work on yourself for yourself, that’s important.
How did you end up in Sweden?
I studied Food Process engineering in Rabat, Morocco. My brother was already doing research in Sweden and he told me about the research opportunities there. So I joined the Master program in Chemistry at Umeå University as an exchange student for 1 year. The university is very good. There was a program for the international students — they gave us an idea of the Swedish culture, helped us to meet new friends, organised parties, and so on, so I didn’t feel like I was alone. In addition, the campus is very nice. The library is the biggest in Norrland [northern part of Sweden]. I still remember spending most of my time there, reading, with my friends. We used to study in a group, and then we had a “fika” and hung out together. It was one of the best experiences of my life. I had my siblings in Umeå, so it was a family thing there — I feel like Umeå is my home city in Sweden 😍
If I say Financial Times, you think…?
The quality of the news is very high. When I wake up in the morning, I like to check briefly the news and have an overview of everything, so it’s such a good newspaper for that. And FT has published a lot of interesting articles about the pandemic during the last year.
Thank you Amina Souihi.
Interested in getting to know the FT? Click here to find out about our Product & Technology teams: roles.ft.com.