Rajinder Tumber: Why did you join the AI industry?
Oana Tifrea-Marciuska: In my case, things evolved naturally. I participated in various competitions during high school that involved math, coding, creative writing, and language. While working on my bachelor thesis, I was able to combine my two passions — language and coding — and worked on “Automatic humor recognition.” This was my first pure hands-on research work into building and interpreting AI models. For months, I would analyse every joke my friends would tell to understand what features made it funny. I continued in this field because I am very curious about various AI-related applications, ranging from its use in education and healthcare to its implications for everyday life, such as helping people find relevant information in various contexts.
Tumber: What do you feel you can contribute to the industry?
Tifrea-Marciuska: I have studied and worked in six different countries, each of which has had distinct ways to approach research and education. When I worked in Bolzano, Italy, some meetings would be trilingual (English, German and Italian), with each participant speaking in the language they felt most comfortable with. My role was to gather the user requirements and make sure there were no misunderstandings. I’ve worked with people from different backgrounds, including educators, linguists, and psychologists. All of these experiences have taught me that many soft skills are necessary in the tech industry.
I see myself as a bridge between different areas of AI, as I have worked on logic-based approaches and statistical models, in addition to being an applied researcher. For instance, when I was working on developing semi-automated games for children who had trouble with reading comprehension, we used various AI techniques to automatically generate games from stories that the children had read. I was continuously working with domain experts (educators, linguists, psychologists), reading not only the technical AI research papers, but also papers from other fields to help me define the requirements. Trials done in both the UK and Italy showed that our system improved the children’s comprehension when reading stories. I continuously aim to give my experience, time, dedication, and focus back to the field.
Tumber: The ratio of men working in AI is significantly greater than women. What do you think women can bring to a male-dominated team/domain?
Tifrea-Marciuska: The only study that I am aware of is the fact that women are better at reading certain kinds of body language than men, which might enhance how they communicate.
Humans are made out of so many different bits and pieces that we inherit from our culture, the books we read, the countries we work or study in, the people we interact with, our experiences, and so on, that I do not want to propagate stereotypes about how men and women are different. I have met men with so-called feminine traits, as well as the other way around. Therefore, I am all for celebrating the diversity of all people, regardless of their gender.
Furthermore, I think that promoting diversity enhances productivity, fosters innovation, and cultivates different perspectives on leadership and collaboration. If we blindly trust the data that we train our models on, we might inadvertently bias our systems toward certain gender or ethnic stereotypes. So we need people from different backgrounds to spot such biases.
Tumber: Being a minority in your team, what (if any) challenges have you encountered?
Tifrea-Marciuska: When I joined Bloomberg, my team consisted of 6 people. As I was the only woman, this raised some self-doubts whether or not I would fit in with the others. Within a very short amount of time, these doubts disappeared. The truth is, every one of the members of my team is actually a minority in some sense. They all come from different countries. Together, we have studied or worked in 13 different countries on subjects ranging from theoretical AI to pure software engineering. To get acclimated with the others on my team, I deliberately arranged informal 1:1 meetings with my peers to get to know them better personally. I believe that having such a diverse team helps Bloomberg build richer products and this also helps me become a better engineer, researcher, and person. What I love about Bloomberg is that its culture promotes the idea of questioning the decisions of others and supports taking risks. It is mainly this culture that helped me overcome my initial concerns.
Tumber: Would you say your career has been a smooth rise to higher positions, or has it had its dips and detours? Please elaborate.
Tifrea-Marciuska: If someone looked at my CV, my career might seem smooth. But you will not find the rejected applications or the uncertainty I had to go through at certain times. Personally, I think that the harder times for me were when unexpected events happened to me or my dear ones, such as serious health problems. For example, I received an important award, but lost someone dear to me in the same week. I feel grateful to have lived and survived some of these experiences because they helped shape me into the person I am today. They taught me about the importance of patience, perseverance, resiliency, plan Bs, empathy, connections, and acceptance.
Tumber: Have you experienced any difficulty with making yourself heard during your career?
Tifrea-Marciuska: It used to be difficult for me to hear myself and trust what I was hearing. I’ve learned that I need to dedicate time to hear to my inner voice in order to understand what I really want and what the obstacles are that stand in my way — and how to overcome them. Otherwise, I might simply parrot the voices of other people as my own.
At the beginning of my career, I was very afraid of public speaking or even showing that I had my own voice or opinion. I went outside my comfort zone by doing a lot of presentations and now I feel better about sharing my own voice.
Of course, I’ve met people that did not listen to what I was saying, but rather made comments that demonstrated their stereotypical beliefs that “You do not look like a computer scientist.” But that is their problem, not mine. What I am always working on is to be a better communicator and this includes both listening and speaking.
Tumber: Within the workplace, do you feel women are being treated differently because they are female?
Tifrea-Marciuska: I think we should encourage that people are treated differently within a workspace. A workspace should provide a safe environment that suits one’s needs, that does not discriminate against others, and that encourages everyone to share their opinions.
At work, I am seen as more than just an engineer or scientist. For example, I am involved with the Bloomberg Women in Technology (BWIT) community, where I have met many role models and listened to interesting talks. I also have gotten involved in many volunteer projects through Bloomberg related to education, nature, or homelessness.
Tumber: Do you think more women should be encouraged to join the AI sector? If so, why?
Tifrea-Marciuska: AI is changing the lives of so many people and we should have a balanced ratio of professionals in the field. There are so many technical and ethical questions to be solved and it would be great to have an equal percentage of female stakeholders in this domain, because many of these decisions might change our lives forever.
The barrier is more of the stereotypical type as opposed to women being less capable than men in tech — as history has clearly shown. Therefore, if we give girls from a young age access to the same opportunities as boys, they will have equal interest and confidence in pursuing this career path. We should not let stereotypes dictate the future of AI.
Tumber: How do you think more women can be encouraged to join?
Tifrea-Marciuska: I think the best thing for children is to be educated without stereotypes about what is a girl or a boy job. Parents should be educated on unconscious bias in parenting the same way we get unconscious bias training at work. When I was in Romania, girls who were good at math were celebrated and encouraged. As a result, I saw a huge number of women studying math, but there were not as many becoming computer scientists — something that was thought to be a career for boys. If young girls (and boys) see examples of successful female engineers, we can change the view of people about what careers are suitable for girls.
For women already working in tech, we need to provide them with tools to keep them there. This was one of the reasons I co-founded the Oxford Women in Computer Science Society (OxWoCS), as I felt we needed more support. This society focuses on bringing speakers and panellists from industry and academia who can serve as role models for the members. We also developed a mentorship scheme and organized various events around networking and personal growth.
For those women who are in non-technical careers, but may want to switch to tech, there are many initiatives, such as Code First: Girls, which provide courses so they can learn how to build their first web application, for example. I have volunteered at many outreach activities and events and have noticed a hunger for this kind of knowledge and support.
Tumber: Do you have any advice to share for female professionals who want to enter the AI industry?
Tifrea-Marciuska: I would advise these female professionals to find or create a support network and determine the area you want to get involved in. Do you want to be more of a research engineer, a data scientist, or something else? Then go and learn the required skills. You can find a lot of free courses available online, including Bloomberg’s “Foundations of Machine Learning” video course.
For people that don’t want to work on the tech side of AI, but are more interested in the ethical side of AI, I would advise them to join forces with AI researchers, as this technology will be touching people’s lives in unfathomable ways.