BACKGROUND & MOTIVATION

Founded by leading experts, Amfileon is a Munich based asset manager focusing on quantitative investment strategies within the Statistical Arbitrage space. 

We use quantitative models based on modern Statistics and AI to uncover inefficiencies in global stock markets and profit from them. Our models are based on a variety of relevant data sets, but most importantly on market data. 

Our trading infrastructure consumes large amounts of market data and evaluates models in real-time resulting in thousands of orders per second during busy market hours. Therefore, it is imperative that we are able to understand, clean and possibly also predict real-time market data.

DESCRIPTION & GOALS

We would like to improve the quality and timeliness of our market data by developing new filtering technologies.

Our system consumes large amounts of real-time market data
- Level 1: top-of-book market data
- Level 2: market by quote
- Level 3: market by order
and will apply newly developed cleaning algorithms based on AI and modern Statistics.

Using the cleaned data as an input we then want to create and store data aggregations. Furthermore, we want to use the cleaned data as an input for new predictive models again based on AI, modern Statistics and basic market laws.

REQUIREMENTS

- Strong programming skills (Python and F#/C# or C++)
- Modern Statistics and AI techniques

REFERENCES

-I. Goodfellow, Y. Bengio, A. Courville, Deep Learning, The MIT Press 2016
-J. Hasbouck, Empirical Market Microstructure, Oxford University Press 2007
-P. Brockwell, R. Davis, Introduction to Time Series Forecasting, Springer 2016
-T. Hastie, R. Tibshirani, J. Friedman, The Elements of Statistical Learning, Springer 2017
-M. Lopez de Prado, Advances in Financial Machine Learning, Wiley 2018
-A. Cartea, S. Jaimungal, J. Penalva, Algorithmic and High-Frequency Trading, Cambridge University Press 2015
-F. Moreno-Pino, A. Arroyo, H. Waldon, X. Dong, A. Cartea, Rough Transformers for Continuous and Efficient Time-Series Modelling, arXiv:2403.10288

Apply to this project here