ÄûÃʵ¼º½

When Chinonyelum Rosemary Igwe searched for projects that connected her passions for machine learning and clinical medicine, she found what seemed to be a perfect match at Aalto University. Moving from an African medical university to a Finnish AI lab, Igwe applied her medical understanding to study how AI can assist decision-making in anaesthesiology.

Perfect synergy of machine learning and anaesthesiology

Igwe is in the sixth year of her bachelor’s studies in Medicine and Surgery (MBBS) degree at the University of Ibadan in Nigeria. Independently, she has also taken courses in machine learning, reinforcement learning, and psychology. For Igwe, the AScI programme provided a rare and perfect synergy. ‘It was the perfect merging of my two deepest interests: machine learning and anaesthesiology,’ she explains. 

To prepare for the 2025 application, Chinonye even wrote an article on Medium about the listed research topic and published a GitHub repository about one of the related algorithms. With this strong application and her unusual but compelling background, she was successfully awarded AScI funding.

Building simulations for anaesthesia

During her internship at Aalto University, Igwe worked on a project titled ‘Simulating medical doctors’ decision-making in anaesthesiology’, supervised by Dr Jan Kompatscher and Professor Antti Oulasvirta in the .

‘Firstly, anaesthesiology is the aspect of medicine where doctors put patients safely to sleep during surgeries and ensure they feel no pain during the procedure. This process involves a variety of decisions concerning the patient’s condition — as every person is different and may be at various levels of health. These include the kind of surgery they are undergoing, which drugs to give and when and how much, when to stop so the patient wakes up at the end of the surgery, and also potential reactions to the drugs have to be monitored throughout,’ Igwe explains. 

Furthermore, her supervisor, Kompatscher, elaborates: ‘Our current work focuses on using Reinforcement Learning (RL) to automate drug administration in anaesthesia. We are focused on finding superhuman policies and, through the analysis of them, finding heuristics for practitioners to apply in real life. To this end, we created a patient simulator on which we can train an RL model.’

This simulator was validated by comparing its behaviour to real-life data and through a Turing-test-like study with anaesthesiologists from HUS Helsinki University Hospital, where they had to tell apart real cases from simulated ones. ‘That Turing-test-like study is 100% handled by Igwe,’ Kompatscher says. ‘Her contribution is extremely valuable, and I am more than impressed with the work she delivers. As mentioned in our previous discussion, I was already very impressed by her application, but now she has managed to surpass the initial expectations I had for this internship.’

For Igwe, the most interesting part was the interdisciplinary fusion of the project. ‘It requires thinking about the way drugs work for patients, how patients are structured, and how they react under different circumstances, and we can reflect these to a good extent in the models we create through maths. That is cool. It challenged my perception of what is possible,’ she says. 

Chinonye truly appreciates the support from the AScI team. ‘I arrived a few days after a very draining examination period of writing papers every day for three weeks for my final-year medical exams. I was excited and exhausted – but hopeful of having a great experience nevertheless. The programme helped me get integrated into the system as fast as possible by providing all the necessary documents ahead of time and providing an informative process on what to do and how. I didn't feel the pressure to hurry or do too many things at once, which really helped me calm down and ease into the rhythm of my new work and environment.’

Igwe’s onboarding was relatively seamless thanks to robust support. Professor Antti Oulasvirta provided high-level guidance, regularly checking in and sharing helpful insights. ‘The lab team members were also very helpful in helping me secure resources,’ Igwe adds. According to her, the group atmosphere was warm and welcoming, and people were happy to help.  ‘We share a good work ethic, and there is a healthy amount of fun. It feels like a safe space where everyone is welcome as they are.’

Her main supervisor, doctoral researcher Jan Kompatscher, organised weekly online meetings to discuss progress and potential challenges while he was travelling for another research project in Tokyo. ‘He was always open to answering any questions I had at any time,’ Igwe shares. Kompatscher adds: ‘Although Igwe received less support than some of the other interns, she had a clear idea of what she needed to do and performed quality work.’

Interestingly, Kompatscher was also an AScI intern in 2024 before he started his doctoral programme at Aalto University. He says: ‘The AScI programme, for me, is a really wonderful thing. Having experienced it now from two different roles—once as an intern and once as a supervisor—I have to say that it is a very beneficial experience for all parties involved. The interns get valuable experience in renowned research labs working on interesting projects. Moreover, the community at Aalto, and especially the other AScI interns, are very welcoming, and strong friendships can emerge from it. Often, a co-authorship or similar relationship can arise from the internship.’

Igwe has advice for those applying for the AScI programme: ‘If you find what you want to do in the projects list, go for it with full force. If it's not in AScI, search everywhere else. If you are not sure, give it your best. If you don't know, find a part of it—or something you do know—and give it everything you can. It's pleasant doing hard work only when it is something you care about, and that increases your chances of becoming great at it.’