For quantitative dev or data scientists at hedge funds, esp. Citadel/2sigma/Jane st, what exactly financial domains are you guys working on? For buy-side, I guess portfolio optimization? What about other stuff like equity research, derivatives or pricing? Or as finance domain layman, what domain knowledge should we pay attention to in order to better equip myself for potential opportunities to work at top hedge funds? I’m MLE/DS super interested in applying ML/NLP stuff into financial fields, that’s why I’m asking for specific sub-domain . Thanks.
Source: got an offer from Citadel and one more ML-heavy market maker. I had a total of > 20 onsite interviews with them combined and only one person asked about fancy things like Transformers, BERT, etc. The majority of ML questions covered linear models, regularization techniques, CV, neural networks. Needless to say, there were a lot of coding questions. Also, I was surprised how many questions were about compression of "deep" models (like knowledge distillation, sort of pruning, etc.).
Huge cong! Your offer is quant or quant developer or MLE at Citadel? How many teams in Citadel play with ML? Thx
This is for Citadel Securities to be clear, market-making business. I do not know how many teams play with ML in Citadel (and Securities), but all the senior people I have spoken to are deeply into some sort of ML (again, not BERT-like, more like linear models).
Hedge funds are not going to be interested in your domain knowledge. Focus on understanding the core concepts well and then applying those concepts to financial applications should be easy.
Thanks. Sounds like funds are still interested in my tech skills like coding and machine learning/NLP?
Coding very much so. ML and NLP have some limited use in the industry. My point is if you know how to code well enough, applying that code to financial problems is a trivial change.