We transfer specific knowledge to help you solve complex problems efficiently.
OUR APPROACH TO CONSULTING
We believe that in many cases, companies that reach out for consulting do not only need a particular, individual solution. They are rather looking for guidance on which approaches to use, which methods to employ, which tools to rely on, and which pitfalls to avoid. We aim to facilitate a lasting knowledge transfer that enables a more efficient problem solving, not just for the problem at hand but also for the ones to come.
Before any agreement, we always discuss if and how we may help you in particular.
We always point towards problem-specific educational resources and courses if available.
We offer focused all-day/week sessions with your team to tackle problems holistically.
Specialized knowledge for specialized problems
Below we specify a few topics that we have specialized on in the past and in which we have gathered experience. If you are not sure whether your problem at hand fits into these subjects, do not hesitate to contact us!
Probabilistic programming represents a very flexible, powerful approach to gain insights even into small data sets with a low signal-to-noise ratio. However, writing probabilistic programs that correctly capture the underlying data generation process can be confusing at first. We guide data scientists through this process, either as a general introduction or for specific problems at hand.
Apart from the modeling process, it is crucial to know how to evaluate the outcome of a probabilistic program, as the inference engine behind a probabilistic programming language may “fail silently” and thus report false insights.
Complex systems are composed from many interacting components. Examples range from power grids and transportation systems, to cancerous tumors and financial markets. The diverse interactions between the components of such a system often results in “emergent” properties such as feedback loops or spontaneous order that cannot be anticipated by looking at the individual components.
Complex systems theory focuses on getting behind these emergent effects, often by investigating the anomalous statistical properties of these systems (fat tails, black swans, bi-modality, …) to better anticipate the future behavior of the systems.
TIME-VARYING PARAMETER MODELS
Time series data generated by complex systems often show regime switches or drifting parameter values that result in anomalous statistical properties such as bi-modal or fat-tailed distributions. Examples are the volatility of financial markets, accident rates in industrial processes, click rates in marketing, cell motility in cancer progression, and many more.
We employ novel methods that not only reconstruct the temporal evolution of the parameters but further allow to identify the type of parameter changes that most likely is behind the anomalous statistics. This type of analysis can be employed both retrospectively and in real-time.
ALGORITHMIC TRADING SYSTEMS
We have experience in the development of automated trading systems for various markets, including US Futures markets and crypto derivative markets. Our focus is on lean code with as little maintenance as possible.
We cover production systems for execution and for risk mitigation, as well as the creation of frameworks for guided strategy development, to shorten iteration cycles and streamline the process from ideation to coding to testing to production.