Discover more from DSBoost
The democratization of AI & finance must go hand in hand! - DSBoost #37
💬 Interview of the week
This week we interviewed Patrick and Janik Tinz, also known as the Tinz Twins.
Studies & Choice
Can you describe your academic background and what led you to choose that particular field of study?
Yes, of course. But first, we'd like to welcome the audience. Hey everyone, we're the Tinz Twins, and our academic journey has been the same. For this reason, please don't be surprised that we always speak of "we". At school, we were very interested in mathematics and physics. In addition, we were also excited about the technological progress of the big tech companies. So we decided to do a Bachelor in Computer Science. During our studies, we learned a lot about software engineering and databases. But very little about machine learning and AI. During our Bachelor's studies, we worked as an assistant in the Mathematics Department. During this time, we had contact with numerous Master's students from the Data Science Master's program. After some discussions, we decided that a Master's in Data Science was the right step for us. We finally completed our master's degree on a part-time basis with the company Accso. Accso is a software consulting company. At Accso, we were able to work on some exciting data science and software projects with a focus on industrial clients. It was a great time. Through our master's degree, we are officially allowed to use the title "Data Scientist". After a year of working full-time at Accso, we decided to start our own business, "Tinz Twins". At Tinz Twins, we focus on Data Science in the financial sector. In addition to our self-employment, we are still studying for an MBA in Finance at the University of Applied Sciences - Burgenland (Austria).
Reflection & Regret
Looking back on your educational journey, is there anything you would change or do differently if given the chance?
No, definitely not. We are very grateful for all the experiences we have had. In our bachelor studies, and at Accso we learned basic software engineering know-how. Finally, in our Master's, we learned analysis skills so that today we can design AI-based applications from scratch. We're just thankful for all the great people we've met along the way so far.
Where do you see yourself in the next 5 years career-wise, and what motivates that vision?
That's a great question. We want to make Tinz Twins a strong brand in the Data Science for Finance field over the next five years. We are currently working on several projects for this. On the one hand, we have our international business line. We teach data science and finance skills in particular. Our platforms for this are X (formerly Twitter) and Tinz Twins Hub. On Tinz Twins Hub, we publish every week exciting articles about Data Science topics. On X, we post educational content about Data Science in an accessible way. Furthermore, we run the Medium publication Towards Finance. The publication aims to build a global investment research community. We want to promote the knowledge exchange between software engineers, data scientists, financial analysts and financial experts. In our opinion, Towards Finance is the place to learn all about Finance, Investment Research, and Artificial Intelligence in the financial sector. On the other hand, we have our German-speaking business. In Germany, we launched a YouTube channel on finance. We also run a weekly e-mail newsletter about finance news. We call this project Tinz Twins Invest. Tinz Twins Invest should be a platform where the community can exchange information about investment topics.
Data Science & Its Trajectory
How do you envision the future of data science, especially with the rapid advancements in AI and machine learning?
AI should improve existing products and workflows. We don't believe that we need an Artificial General Intelligence (AGI). We need specific AIs that add value in specific domains. Yes, ChatGPT is a great tool to solve rudimentary tasks. But no more than that. It doesn't work well when we want to solve specific tasks. It doesn't help to use retrieval-augmented generation (RAG) either. At some point general models reach their limits, so we think we should rather use fine-tuned models for specific tasks. Apple and Meta go this way. Both are trying to make their products better through AI. That's the right way to go, in our opinion.
Career Alignment with Data Science
Given the current trends in data science, how do you plan to align your career with these developments?
The performance of today's large language models (LLM) is impressive. In the future, we'd like to focus more on LLMs in the financial sector. We think that LLMs have the potential to change many industries from the ground up. For this reason, we have been planning to develop an LLM-powered investment research app for some time. The app is currently under development. If you want to stay informed about this project, follow us on our channels.
Data Science Impact on Finance
In your opinion, how might the evolution of data science affect the finance sector in the coming years?
In our view, LLMs will play a big role in the financial industry, the so called FinLLMs. Examples are BloombergGPT (a proprietary LLM for finance developed by Bloomberg) or FinGPT (developed by the open-source community AI4Finance Foundation). We think the future of FinLLMs and financial platforms should be open-source. A great project in our eyes is OpenBB. OpenBB is an open-source investment research software platform. The platform gives you access to high-quality financial market data and analytical tools. OpenBB's goal is to make investment research accessible to everyone. In addition, OpenBB is already beginning to integrate LLM-powered features into its platform. OpenBB Terminal is free. A license for a Bloomberg Terminal costs $24,000 per year.
Considering potential shifts in data science that could impact finance, how do you plan to adapt or evolve in your role?
In the future, we want to focus primarily on FinLLMs. We also want to educate ourselves further in the area of finance. Our vision is to make Data Science for Finance accessible to everyone. That's also the reason why we decided to do an MBA in Finance. We also love to learn new things and invite everyone to become a Towards Finance writer. On Towards Finance, we want to share our knowledge about finance and data science with everyone in the future. Let’s bring together the know-how of all the experts in these fields because we can achieve great things together.
Skills & Knowledge
What new skills or areas of knowledge do you believe will be crucial for professionals in the intersection of data science and finance?
In the financial industry, many products are proprietary. In the future, it is essential to take a look at open-source products. It happens a lot in the field of open finance. Soon, Bloomberg's high prices will no longer be justified as there will be cheaper and open-source alternatives. The use of LLMs in the financial industry will keep many professionals busy. How can we use LLMs efficiently in finance apps? How can we fine-tune LLMs for financial use cases? It is an interesting and exciting topic for the coming years. The democratization of AI and finance must go hand in hand!
For someone looking to enter the world of data science with an interest in finance, what advice would you offer?
It is essential to have an understanding of time series analysis. In addition, you should learn how time series algorithms (RNNs, LSTMs, GRUs, ARIMA, ...) work. If you need data for exploration, we recommend using the OpenBB SDK. We have learned one thing in the last few years. It is super important to understand the specific domain. A good data scientist has domain-specific knowledge, which allows him to generate added value from data. A deep understanding of the task is essential, and you should keep the task as small as possible. And you'll learn the most by working with real-world data. Also, you don't necessarily need the largest ML model to solve tasks. Simple models are often sufficient to achieve good results.
How do you think data science could bring about interdisciplinary collaborations, especially in areas like finance, and what benefits do you foresee from such collaborations?
Thanks for this interesting question! An exchange with financial experts is essential to use data science effectively in the financial sector. We think it's important to exchange ideas with others from different domains. For this reason, we want to promote this exchange through our Medium publication Towards Finance. Only together can we achieve great things. Data science, financial knowledge, and technological progress must go hand in hand. That's how we can make the financial industry more transparent, accessible, and efficient.