infinada.com - Consulting Services

Consulting Services

infinada Consulting GmbH provides consulting, coaching and training services in the following areas:

Statistics, Data Analysis & Data Science

analysis of data, including large collections of unstructured data ("NoSQL")
application of statistics and machine learning, data analysis with R ("data science")
forecasting and classification ("business analytics")

Valuation and Risk Models for Complex Financial Products

valuation of complex financial products (e.g. financial derivatives) and insurance liabilities (e.g. variable annuities)
portfolio optimization and asset-liability management (ALM)
risk models covering market risk, counterparty credit risk, insurance risks (life, health, non-life) and operational risks

Actuarial Consulting

actuarial analyses of life and health experience data, supporting product design, pricing, reserving and risk modeling
creation and improvement of actuarial assumptions for pricing and reserving
supporting the implementation of new products in policy administration systems
business analyst tasks, like interface management, business specification and test management
optimization of reinsurance cover

Independent Model Validation

independent model validation of risk models in the context of Solvency II
independent validation of credit risk and market risk models in the context of Basel III and IV

Strategy Consulting & Coaching

coaching one or more IT teams (e.g. as Scrum master)
project management in traditional and agile projects
entity relationship modeling
consulting on data strategy, risk strategy, business processes and organizational structure
how to combine agile techniques (e.g. Scrum, Kanban, continuous deployment) with longer-term goals
mentoring

Data Analysis Tools

NoSQL databases (like Redis and MongoDB) are powerful tools allowing to process fast-flowing, unstructured data.

R has been the leading open source statistical software package since 1997. It is one of the two dominant analytics languages in the "big data" and "machine learning" context besides Python.

Complex Financial Products

Life insurance and pension contracts (e.g. variable annuities) are among the most complex financial products. They are a challenge to properly value and risk-manage due to (1) their long-term nature, (2) the multitude of product parameters and (3) the optionality originating from the options for the policyholder to adjust the contract parameters and for the insurance company to adjust the management of the portfolio.

Independent Model Validation and Quality Assurance

Concerns have been voiced that future systems employing "machine learning" might be so complex that humans will be unable to test or validate their reliability. The key to maintaining control over complex systems is to tell the difference between models which can be validated and models which cannot be validated.

On the one hand, there is a class of surprisingly complex machine-models (e.g. for near-optimal strategies in games like Go or Chess), which can be efficiently validated by a human equipped with a (single) classical computer and a source of "true" randomness. (Even though finding the near-optimal strategy is way beyond the time and energy constraints of a single human with one computer.)

On the other hand, there are seemingly simple machines (e.g. Pokemon Yellow), which are "accidentally Turing-complete". And it can be proven that very simply questions about their behavior is undecidable, let alone that desirable performance characteristics can be guaranteed.

A very closely related insight is that the language used to ask machines to perform certain tasks should be suitably restricted, e.g. to regular languages. Otherwise it is difficult to provide security guarantees for those machines. (And no security guarantee whatsoever can be given for machines which need to interpret languages that are Turing-complete.)

For the full story see the February 2020 essay "Why model validation is feasible".

Management

Scrum is currently the most popular agile software development method.

Kanban is a method of increasing throughput and quality in business processes, originally developed by Taiichi Ohno at Toyota. The main principles setting it apart from other "lean" methods are (1) strict adherence to work-in-progress limits, (2) later process steps pull work items instead of earlier process steps pushing them and (3) in case of a bottleneck the whole process chain is stopped and the team joins forces to remove the bottleneck.