What I like most about my job
What I like most about working at Ortec Finance are the competent and skilled people. It's good for your personal learning curve to be surrounded by people with superior knowledge. As the work is rather challenging, these people can serve as guides within a complex learning environment.
What I did not expect to learn that fast
I was always a talented programmer at my Mathematics and Econometric studies. However, working in a professional software development team I quickly realized that I was an amateur.
When you're thrown in the deep, you sink or swim, and it surprised me how fast I learned how to swim and how to effectively contribute to the team. Learning the basics of new coding languages, workflows, and all kinds of tools within a matter of weeks, which were all previously new to me.
Student Assistant/Software Developer for GLASS
My role as student assistant
At Ortec Finance, I am researching the interpretability of reinforcement learning for asset allocation. This is a challenging subject, as we are dealing with a continuous and high-dimensional action & state space. The subject perfectly matches my background in Mathematics and Econometrics. Ortec Finance grants me total freedom in the research direction, which suits me best. More on my reinforcement learning project below.
Besides this, I am working as part-time software developer for one of our solutions; GLASS. This is challenging in a totally different way, my team is made up of exceptional developers, whereas I have mostly been coding as a hobby before this. I was passionate about my desire to strengthen my programming skills, thus I currently am focusing on being a real contributor to the team.
Again, there is significant freedom and responsibility paired with this job, which is what makes you feel part of the team instead of just some assistant.
Given Ortec Finance's commitment to innovation, I foresee a future for myself within the company. Reinforcement learning and other machine learning architectures will get more important and within Ortec Finance I can contribute to this.
About my main project – Reinforcement learning
Econometric models based on mathematical optimization have been the most popular way to guide large financial decisions. Through their simplicity, these models grant high-level insights and interpretability. However, they lack flexibility and generalization performance. With Ortec finance, we are training "agents" (neural networks) in simulated scenarios to make investment decisions.
This training is done with reinforcement learning, where we combine powerful computing with state-of-the-art algorithms to train agents which are both extremely flexible in their objectives and have high performance. Of course, we lose the inherent interpretability we had with legacy methods, but it is my personal belief that performance and interpretability are not orthogonal to each other. Retaining a lot of performance and increasing the interpretability significantly can hopefully be achieved simultaneously, which is the objective of my project.
The R&D Labs
Ortec Finance puts innovation through applied scientific research and IT developments at the forefront. Its research and innovation activities are concentrated within their R&D Labs. Next to projects ran by permanent staff, they collaborate closely with universities by supervising about a dozen Bachelor, Master, and Ph.D. projects per year.
Interested in joining Ortec Finance?
Did the story of Gert trigger your enthusiasm for starting your career at Ortec Finance? Check out our vacancies!