You will spend a lot of time writing near-production quality code, so you have to be very comfortable with Java or C++, as well as advanced R, Python.
You will enhance and maintain portfolio risk budgeting, risk measurement infrastructure and be responsible for the integrity of the underlying quant library.
You will spend a lot of time thinking about the limitations of the machine learning models and human-machine interaction policies.
You should be comfortable with the role of a devil's advocate, as your primary responsibility will be to expose weaknesses in the models and communicate them clearly and openly.
Advanced degree from a top program in Mathematical Finance, Applied Mathematics, Electrical Engineering, Statistics or other related fields. PhD preferred, but an exceptional candidate with MS and a strong industry experience may be sufficient.
Formal training in data science is not necessary, but as a bare minimum you should have done an online training course (i.e. Coursera) or data science course at a coding school.
Experience modeling credit instruments and portfolios is a must.
Must have humility about the limits of one's knowledge.