WoW Woman in Health Tech I Magda Dubois, Data Scientist at TORTUS
Magda Dubois is a Data Scientist at TORTUS. She is a computational cognitive neuroscientist (UCL PhD) and biomedical engineer (Imperial College MSc, EPFL BSc) who was trained & educated in London and Switzerland. Her PhD research focused on the computational modelling of human decision-making and its link to mental health. She has carried out research in Boston (Harvard Medical School), Tokyo and Germany, and she took part in the Entrepreneur First programme in 2023.
TORTUS is a London-based startup founded by Dr Dom Pimenta and Christopher Tan in the prestigious Entrepreneur First programme. TORTUS was the UK's first ambient voice technology company to achieve NHS assurances and is currently used in NHS hospitals and GP surgeries across the UK. They are backed by Khosla Ventures and other notable investors, including former chair of the NHS, Lord David Prior, former NHS COO, Sir David Sloman, former director of machine learning at X, Rob Bishop, and VP of AI at 1X Robotics, Eric Jang.
Tell us a bit about your background and your projects so far.
I studied biomedical engineering, earning my Bachelor's at the Swiss Federal Institute of Technology Lausanne (EPFL) and my Master's at Imperial College London, where I specialised in neuroscience. I spent two years doing research in labs across the US, Japan and Germany before moving on to complete a PhD in computational cognitive neuroscience at UCL. After that, I decided to transition into the industry, so I joined a data science placement programme (S2DS) and participated in the Entrepreneur First start-up accelerator to get hands-on experience in the startup world. Currently, I’m working as a Data Scientist at TORTUS, focusing on evaluating large language models (LLMs) for healthcare applications.
How did you get into this industry? Has it been an easy industry to get into, or have you had many challenges?
I built a strong foundation in scientific methods, data analysis, and statistics during my studies and research in academia, particularly at the intersection of neuroscience and technology. When I decided to move from academia to industry, I worked a few months for a neurotechnology start-up, completed a data science placement and then joined the Entrepreneur First start-up accelerator, which was like a hands-on mini MBA.
The biggest challenge for me was making the transition from academia to industry, especially in the start-up world. While the methods are similar, the vocabulary, goals, pace, and communication styles are quite different, particularly when working with non-academic teams. Adjusting to this new environment is a work in progress, but it is a rewarding journey!
How long did it take you to be where you are now? What was the biggest obstacle? What are the challenges of being in the industry you are in?
As previously mentioned, the biggest challenge for me has been transitioning from the academic world to the industry, as it took some time to adapt to the startup environment.
As for the field, I'm currently in—AI in healthcare—it presents its own set of challenges. Since AI models rely on data rather than explicit programming, there is a lack of transparency in how they operate. We often don't fully understand the inner workings of the models and can only rely on evaluating their past outputs to gauge their capabilities. Given how rapidly the field of LLMs is evolving, standardised evaluation methods are lacking and we need to continuously evaluate what is coming out. This is specifically difficult in healthcare as we face the added responsibility of patient safety.
Another related major issue is the need for vast amounts of clinical data to train or assess these models effectively, which is often not an option due to obvious privacy concerns. Curating meaningful, representative clinical datasets—whether anonymized or synthetic—is an intensive and time-consuming process. Addressing these challenges at TORTUS is demanding, but it's also incredibly rewarding and exciting work.
What are your biggest achievements to date?
One of my biggest achievements at TORTUS to date has been co-leading the development of a platform specifically designed for the evaluation of AI-generated medical notes by clinicians. This not only helps us build a large, valuable database of medical notes that can be used to train and fine-tune AI models, but it also serves as a comprehensive evaluation framework. We use it to assess our various models and metrics, helping us continuously improve the quality and accuracy of AI-generated medical notes.
What are the projects you are currently working on?
Currently, I'm working on several exciting projects. In AI evaluation research, I'm focused on developing and maintaining a human annotation pipeline, which involves creating metrics and processes to ensure that AI-driven tools are safely integrated into clinical settings.
Within clinical research, I'm helping design and statistically analyse multi-site clinical trials in collaboration with the NHS. Additionally, I'm involved in user research, where I create visualisations and analyse user behaviour and patterns to help shape and refine our product and commercial strategies.
Is the #WomenInTech movement important to you and if yes, why?
I’m deeply passionate about the #WomenInTech movement. While it is encouraging to see gender gaps gradually narrowing, there is still much work to be done. Movements like #WomenInTech are vital because they amplify the stories of women in the field, foster networking, and create spaces where women can connect and communicate more freely. One consistent finding across studies in cognitive neuroscience is that women (vs men) tend to exhibit less confidence, even though there is no difference in actual performance. Women-focused events offer environments where they can express themselves more comfortably and connect in ways that feel natural.
What will be the key trends in your industry in the next five years and where do you see them heading?
In the next five years, I believe a key trend in the AI industry will be the growing regulation and oversight of AI models. Currently, AI models can be used with relatively few restrictions, but with initiatives like the EU AI Act and the establishment of global AI safety institutes (e.g., the UK AI Safety Institute), we can expect more structured guidelines and regulations. While deploying models in different applications may take longer, these measures will significantly enhance safety.
A big part of my work at TORTUS is to define and document best practices to ensure patient safety and to be prepared for when these regulatory changes come into effect.
What is the most important piece of advice you could give to anyone who wants to start a career in this industry?
My key advice for anyone starting out in this industry is to focus on developing strong communication skills. If you are interested in data science, you likely already have the technical know-how, like coding and statistics, but the real challenge is explaining data insights to people who may not be familiar with them. I don’t have a perfect solution to this, but investing time in learning how to do this effectively will make a significant difference in your career.
Who are three inspirational women in your respective industry that you admire?
There are many women in the industry I admire, but to name just three:
Yael Niv - a computational cognitive neuroscientist and professor at Princeton University. She is among the most cited cognitive neuroscientists in the world. Her work has been instrumental in advancing our understanding of reinforcement learning in the human brain and she was among the first to connect these computational theories to mental health and psychiatric disorders. She is also a strong feminist advocate and pushes for greater equity and diversity in academia.
Amanda Cox - editor at The New York Times and an inspiring leader in data visualisations. She creates some of the most insightful visualisations in journalism, making even the most complex data easy to grasp at a glance. Turning messy, intricate datasets into clear, intuitive charts requires a rare combination of analytical, creative and communication skills. She has a unique talent for distilling key messages and presenting them in ways that resonate with diverse audiences.
Ida Tin - co-founder and CEO at Clue. Not only did she bring global attention to women’s health by coining the term “femtech,” but she also developed a groundbreaking app in this field. I’m a big fan of Clue—not just for its approach to period tracking but also for its commitment to using this data to advance scientific research and enhance our understanding of women’s health. They also offer a wealth of educational resources that help people gain a deeper insight into women’s bodies. Given the significant data gaps in women’s health, initiatives like these are crucial for driving meaningful progress and making a real impact in the field.
If you want to share your story with us and be featured on our WoW Women blog, get in touch.
This interview was conducted by Marija Butkovic, Digital Marketing and PR strategist, founder, and CEO of Women of Wearables. She regularly writes and speaks on topics of wearable tech, fashion tech, IoT, entrepreneurship, and diversity. Connect with Marija on LinkedIn.