Project volunteer Heiko Hotz
“DataKind is looking for people who are passionate about data science and willing to learn.”
Pronouns He/him
Roles AI/ML Solutions Architect at AWS, Freelance Consultant at AI/ML Consulting, Founder of NLP London
Data Ambassador at DataKind UK
Heiko has been a volunteer with DataKind UK for several years, including as a Data Ambassador leading a data science project with anti-corruption campaigner Global Witness.
Tell us a bit about yourself and your data background
I’m passionate about building solutions with AI and Machine Learning (ML) and sharing my experiences through blog posts or at public meetups and conferences. Over the past 20 years, I have always worked in IT in one capacity or another. I believe if you’re passionate about a particular topic, it is best to pick up the skills as you go, rather than waiting until you think you’re “good enough”.
Why did you decide to become a Data Ambassador (DA)?
In 2016, I was a director in the data analytics group of a consulting firm and started getting more into data science. I was really intrigued by the possibilities that ML offered, and took a few online courses. My day-to-day work didn’t offer many possibilities to practise these skills, so I looked for opportunities elsewhere.
That’s when I found DataKind UK online, and thought this would be a great opportunity to donate my time and skills, and hone my data science skills. And when I started speaking with the people at DataKind, I quickly realised that they are the kind of people I want to hang out with!
What did you enjoy the most about being a Data Ambassador, and what did you find challenging?
By far what I enjoy most as a DA is that I get to know many interesting people from completely different backgrounds. It was absolutely fascinating, for example, to work with Global Witness, an organisation that investigates and exposes environmental and human rights abuses in the oil, gas, mining, and timber sectors, and tracks ill-gotten money and influence through the global financial and political system. To see how they use data and how, together, we can use ML and data science to help further their mission was one of the highlights of my DA journey.
What advice would you offer to future Data Ambassadors?
Above all, I would encourage anyone who is interested to apply, no matter where you think you are in your journey. In my experience, DataKind is looking for people who are passionate about data science and willing to learn, more than anything else. Even if you are not sure your skills are “good enough”, I’d recommend reaching out to the DataKind team.
What surprised you most about volunteering?
I was surprised by how many different backgrounds the other volunteers at DataKind have. But in the end, this makes total sense — data science is such a creative, multi-faceted discipline, and it greatly benefits from many diverse perspectives.
What is your daily data work like?
I work with customers to adopt AI/ML in their organisation. This could mean that I review their architecture and provide guidance on best practices or build a demo that showcases the art of the possible in AI/ML. More often than not the use cases are cloud-based, and as such, I often use Amazon SageMaker as the ML platform of choice to train and deploy models.
What do you wish you’d known when you started to get into data?
I can’t say I didn’t know this before, but I think it bears repeating that time spent on improving data quality trumps time spent on improving ML models every time. We see this in the space of large language models like GPT-3 — having a better-curated dataset trumps sheer model size. This is a trend that I believe will become even more important in 2023.
Did you always think you were going to go into data?
I think I always knew that I would end up somewhere in the IT landscape. At some point, I considered going into the direction of Chief Information or Technology Officer of an organisation, but I realised I love building stuff and being hands-on too much to be a strategic leader. So now I focus on building solutions and becoming a subject matter expert in AI.
What is a data project that inspires you?
Not so much a data project, but the open-source AI project Transformers from HuggingFace. It allows unprecedented ease of use for some powerful AI models. My customers are fascinated by how easy it is to leverage state-of-the-art Natural Language Processing with this library.
Is there a resource you’d recommend to the community?
If you would like to get into deep learning, which is a good field for getting started in data science, I highly recommend the (free) course Practical Deep Learning by FastAI.
Tell us something completely non-data-related about yourself!
I love travelling, hiking, and food — and the best thing being they are very easy to combine!
My recommendation for two of the best hikes I have done in my life are the Inca Trail in Peru and the West Coast Trail in Canada.
