In-depth project – S1


Students in the M2 "Statistics, Finance, Actuarial Science" and the Actuarial Science and Finance & Risk Management tracks of the ENSAE engineering cycle have the opportunity to participate in a prospective work, in groups of 3 or 4 and under the supervision of a supervisor, financial industry professional or researcher in finance/insurance. The supervisor will meet the students once or twice a month to coordinate and guide their work, in person or by videoconference. He or she will also provide the students with the data necessary to conduct the project.

In-depth projects focus on a well-identified topic in quantitative finance or risk management, and have a real interest in the company. For example, it can be an exploratory research around a new model, the analysis of a specific database or the study and development of a new management strategy. Projects will include typically a bibliographic research stage, a data analysis stage and a computer implementation stage at based on algorithms. Although the results of the project may be interesting and useful for the company, it is by no means a consulting mission.
The objective is to be prospective, to test innovative ideas.

The publication of an article, the result of such work, could be considered in some cases.

An intermediary report will be issued at the end of January, a final report in May and an oral presentation. This is a fully-fledged course, which will allow the validation of 3 ECTS in the first semester and 3 ECTS in the second semester.

Examples of topics proposed in 2019-2020 :
– Generative Adversarial Networks for Stress Tests (Natixis)
– Valuation of securitization products (Banque de France)
– Modeling of credit card balance (Banque de France)
– Initial coin offerings (PWC)
– Dynamic allocation across S&P500 sectors (MarketCipher)
– Reinforcement learning for asset-liability management (EDF/Amundi)
– Construction of statistical tests for martingales and application to the validation of risk-neutral economic scenarios (Milliman)
– Machine learning in portfolio management (BFT Investment Managers)
– Advanced arbitrage-free volatility and price surfaces (Zeliade Systems)
– Statistical analysis of the electricity intraday market (EDF R&D)
– Prediction for the optimization of trading strategies on the electricity markets (CREST).


The indicative timetable for the implementation of the projects is as follows:
– Beginning of October: dissemination of the topics to the students,
   group building
– End of January: submission of the mid-term report
– Early May: submission of the final report
– Before the end of May: defense

This training aims in particular to give students the following skills:
– Bibliographical research
– Analysis of financial data
– Formulation and implementation of a mathematical model based on an 
business issues
– Methodology, specificity and requirements of industrial/academic research