Last week, we brought together four of our exceptional colleagues for a panel discussion on careers in machine learning applied drug discovery. Together, they explored what life is like in a fast-moving, multi-disciplinary and collaborative scientific environment and challenged some of the popular misconceptions about working in this industry. Here are some of our main takeaways:
1. Be willing to tackle challenging problems and embrace new ideas
Human biology is immensely complex, and the application of machine learning to drug discovery is an accordingly complex challenge. To come up with solutions, we need to combine a variety of scientific and technical skills and backgrounds. Ultimately, no one has all the answers, but each discipline brings valuable insights. For technologists to be successful in this field, therefore, they do not necessarily need a scientific background, but they need to be curious about domains beyond their own and be willing to learn.
That is not to say that the learning curve won’t be steep, but anyone open to new ideas and approaches can flourish. As Aylin Cakiroglu, Senior AI Scientist, said, “My first year was really difficult, but I remedied that by asking plenty of questions – and kept asking and asking! Thankfully, I had fantastic colleagues who never made me feel stupid for doing so.”
Beyond asking questions, there are so many resources out there that you can learn from; our team referenced using Coursera, reading books, and watching lectures online. At Benevolent, we have also created internal resources to help get new joiners up to speed, as well as learning and development programmes. You won’t be alone.
Paidi Creed, Director AI Science, described his bemusement at how his knowledge of biology has progressed during his time at BenevolentAI, “From a fair amount of self-study and working on projects, I could probably bluff my way through the coffee break at a biomathematics conference!”
2. Collaborate, communicate and respect the expertise of others
As Saee Paliwal, our Lead AI Scientist says, “There is no cookie-cutter of the perfect employee, and we need all hands on deck to solve the problems we’re tackling.” Many of the assumptions that led to success in a particular subfield may not hold in an interdisciplinary environment: you often need to respect and learn from the expertise of others.
Every subfield also has its own terminology, and it takes work and commitment to create a common language that allows people to understand each other. In Aylin’s experience, creating successful multidisciplinary teams is “an active, ongoing process.” While it may be difficult, breaking these boundaries leads to innovation.
Saee believes the best way to enable collaboration is to align on the end goal: accelerating drug discovery. That way, you can tell people: “Leave your biases and your egos behind, it’ll be worth it.”
3. Don’t let imposter syndrome stifle your potential
Science and technology are notably behind the curve in terms of welcoming and celebrating people from diverse backgrounds, and sadly there is still considerable work to be done. It can be challenging to be a woman in a male-dominated field. Saee often felt her voice was disregarded by others during her studies and early career, which made her doubt her abilities; “The more others disregard you, the more you disregard your voice internally.”
Imposter syndrome can cause women, and underrepresented populations more generally, to underestimate or undersell their capabilities, which risks stifling potential. Aylin recounted how she did not believe she had the skills for a full-time role when she transitioned from academia to industry and was going to apply for an internship at Benevolent until she was persuaded otherwise. Indeed, at BenevolentAI, we encourage potential applicants not only to apply for jobs where they fit 100% of the requirements; sometimes just 50% is enough. As Ana Leite, Lead Bioinformatics Data Scientist, summarised; “We look for talent and potential. We just have to be confident that you are capable of learning the other 50%.”
4. Challenging problems will be solved with diversity and inclusion
Diversity inspires innovation and inclusion shapes culture. Combined, they drive business success, which for Benevolent means better new medicines for patients. Ana underlined the importance of diversity at Benevolent, where “we want to solve previously unsolved problems and so we have to think differently. The only way to have this innovative spirit and come up with new approaches [...] is to have a really diverse workforce. This won’t happen if everyone is thinking the same way.”
Beyond hiring people from diverse backgrounds, Benevolent is committed to ensuring everyone feels a sense of belonging, respect and value. Reflecting on her own struggles with imposter syndrome, Saee identified how this inclusive environment allowed her to thrive and contribute her best work; “Thankfully, Benevolent is an incredibly supportive, inclusive environment, and working with such wonderful colleagues has really helped me overcome my own self-doubt.”
Supporting employees throughout their career also means allowing working life to evolve as family life changes. Ana, who has three small children, commented: “Maternity and paternity leave cannot be a blocker in terms of career progression: it’s part of life and we need to be able to progress with or without taking it.” Paidi argued that offering greater support and encouragement for men to parental leave can help to reduce glass ceilings for women in work. Indeed, when his daughter was born, Paidi was able to restructure his role at Benevolent to allow him to both care for his child and support and enable his wife’s career as a Professor.
5. When we breakthrough, it will be life-changing
Without exception, our panel spoke with huge excitement and optimism about the future of AI in drug discovery. The quality of data available, methods used, and modalities are improving every day, enabling better and more sophisticated research and hypotheses. Paidi mentioned the increase in large, publicly-available data sets such as UK Biobank and Human Cell Atlas and the potential to leverage these to build more complex machine learning models.
Ana also stressed the ongoing need for quality data:“There is a lot of potential, but we don’t only need a lot of data - we need good data. This is what Benevolent is working on: generating data that is suitable for AI and ML.” There will be challenges ahead, but when we do breakthrough, it will be life-changing.
If you are interested in adding your skills to the mix and helping in our mission to develop life-changing medicines, then check our job page and/or internship page for more information.
Back to blog post and videos
