Meet Tonia. She's Director of AI and Analytics Centre of Excellence.

Can Chat GPT do a sauté?

The goal is to bring data to life using AI and Machine Learning, convert data into something that is useful to the business and, ultimately, improve patients' lives. This is what I do at the AI and Analytics Center of Excellence, which is a unit in Novo Nordisk Digital Data and IT. Here, we are a group of data scientists, machine learning engineers, and software developers who work across the organisation. We partner with data scientists and business experts in different projects throughout the organisation, all the way from R&D to clinical trials to production to commercial units.

We also have an MLOPs team with the responsibility of creating services to reduce time to market for the deployment of AI algorithms. Additionally, we have an AI lab that focuses on new AI technologies, such as generative AI, and explores how we can adopt relevant AI innovations, bringing them in the company in a fast and secure manner.

Pas de bourrée

✦ I really like dancing, I used to be a semi professional ballet dancer. I even went to a professional ballet school, when I was young. Not really what you’d expect from an AI nerd.

Joining

Ultimately, everything is data

To be honest, before joining, I was a little bit afraid. I knew all the good things about the company – the purpose, the commitment, the culture, the scale. However, given the scale and the size of the company, I thought it might be a little bureaucratic, slightly low-tech, a little slow – partly because it's a regulated company, obviously for good reasons. 

But my unit was moving way faster than I expected. It really had this kind of start-up mindset, and the people in it wanted to change and improve things. In general, everybody in Novo Nordisk is really committed to making this company a better place. Everybody is very driven and competent. Because you work with very committed people, it also helps you improve, as you're learning from all those good colleagues. I feel positively impacted by my colleagues’ energy and drive.

I also decided to work for Novo Nordisk because of its purpose. Ultimately, everything involves data, but when you know that the data you work with, and the programs you create, ultimately contribute to improving a patient's life, it's highly rewarding. This alignment of personal and company purpose is a significant motivator for me. And the company culture was another crucial factor in my choice. Despite being a global company, it upholds inherent Scandinavian values, emphasizing equality, respect for employees, environmental responsibilities and ample opportunities.

Work

Let’s improve the prompt guide

My team does some very cool projects. My work is to set the direction together with the team and most importantly support and empower my teammates doing all this cool tech work J  As far as I know, we were one of the first companies in Denmark and probably the few globally from our industry and size to adapt ChatGPT-like services internally so quickly. Early spring 2023, we created the Novo Nordisk ChatGPT allowing everybody in the company to start using it, innovate with GenAI while ensuring safety. We have thousands of daily users and we are also constantly adding new general features that can help all our 60,000-ish colleagues, as well as working with special LLM applications with internal data. Our users span across the company; from medical writers, to production operators, clinical trial experts, R&D scientists, commercial colleagues and sales reps, for example. So that's amazing to see.

Tech

The one with the API keys

Well, nowadays I mostly do Power Points and strategy about AI :-) However, my team primarily operates in the cloud with AWS and Azure. We follow a dual-cloud strategy and utilize tools like SageMaker and Databricks. We also work with code, data and ML engineering pipelines to construct various AI models. This typically involves using Python, and sometimes we serve our models via APIs. We provide ML Ops services and leverage Azure DevOps and Github extensively, frequently employing continuous integration and continuous deployment (CICD) practices. Furthermore, we constantly strive to enhance our best practices to make these services and models more accessible to users. We achieve this by developing custom applications, such as Streamlit or Dashboards or React apps, and, naturally, by offering these models through APIs that can be integrated into various downstream applications. 

Challenge

AI will give you time

I know it's gonna sound cheesy, but I do think my biggest challenge is that there's so many things to be done. What is the potential of using even more AI? How can we empower everybody in the company by using these tools in their daily work? Or how much can we help other data scientists scale AI with infrastructure? There's almost not enough time. But that's a luxury challenge. It’s a massive company with massive amount of data in multiple modalities and business domains. How can we become more AI-driven by AI and when necessary, how can AI become a daily tool, both in our business decisions and in every employee's life? That’s an interesting challenge with enormous potential. 

Communities

Master and apprentice

Internally, we have a data science community that I’m, of course, part of. I also have awesome teammates and leadership colleagues which I continuously learn from. 

Externally, I am also following my field. I go to data engineering or data science meet-ups, and regularly attend or present at conferences. Last year, I was also fortunate enough to be recognised as AI Professional of the Year in the Nordics by Hyperight, so I obviously try to contribute to this community as well. I do my best to support new female employees, and mentoring when I can, and I would like to do more of that. I’m also supporting the Women in AI/Tech communities in Denmark. 

Next up… In search of lost time

🕮 I am in a book club. I really like literature as a way to force myself to also use the other side of my brain - not only the engineering one.