XVA calculations module is a distributed module responsible for scientific computations of financial risk measures: XVAs, as well XVA Greeks.

Operating similarly to MapReduce algorithms, this module generates large data volumes on a hundred of computation nodes, that are then “aggregated” (reduced, in MapReduce’s terminology) into XVAs. The system generates large reports (nightly runs, sometimes requiring hours of computations) as well as incremental computations (daily activity).

The performance of the Aggregation step is critical to our clients’ XVA desk activities. The part

of the module has evolved in the past years due to:

·       Introduction of distributed aggregation,

·       Needs to compute Greeks during the day,

·       New type of deployment architecture: Amazon AWS.

Missions:

The goals of this internship are:

1.      to extend the existing tooling for performance testing: automation of runs on large data sets, automation of the generation of reports, distributed aggregation coverage,

2.      to develop performance improvements thanks to the tooling (proving its value),

3.      to ease the adoption of the added capabilities by the team, thanks to demos and documentation.

Master’s degree in computer science.

« Mandatory » skills:

·       java development: tools and production code,

·       CI (Jenkins) development: automation,

·       software optimization basics.

« Nice to have » skills:

·       Distributed computations,

·       Financial computations: XVAs.

Resources

https://en.wikipedia.org/wiki/XVAhttps://en.wikipedia.org/wiki/Greeks_(finance)https://en.wikipedia.org/wiki/MapReduce

Duration: 6 months from March/April