I have recently started posting on blind and I must say I have learned alot. Some really true stuff is projected here which someone like me would have kept guessing if I was not here. So this one goes out to all the DS hiring managers. I recently interviewed with two of FAANG companies and got rejected. So I was just doing self analysis of where things went wrong. For someone who is a DS HM how would you guage a candidates skill-set and make a descision on them. I was wondering how much weight is given during an interview to the following: 1) understanding of the basic ML algos (RF, KNN, ARIMA, Regression etc) 2) Discussion around model evaluation metrics 3) Medium- hard level statistics (MLE, p-values, hypothesis testing, distributions etc) 4) Understanding of business and/or prior experience in that particular area. 5) Desining experiments (A/BTesting) 6) Model productionizing (a touch of SWE) 7) Communicating the findings to decision makers ( how would you explain regression coefficients to a PM or do you even need to?) 8) coming from a PhD background 9) Databases (SQL, Postgres etc) In nutshell at what point you think that there are enough signals to bring someone in? Thanks
It’s usually a team that votes, and not necessarily a hiring manager. I’d say team/culture fit plays a significant role. On the technical side, just make sure you know the fundamentals and know them well. I’ve interviewed candidates who claimed to have used a programming language, but couldn’t write a simple for loop! Resumes are usually exaggerated, not a good sign! Two rejections don’t mean much. You should post if you get 20 rejections. Stay positive and good luck.
I have received 20 rejections, and I can write for loops. The DS domain is so unstructured that it's difficult to nail down one reason for the rejections.
After 20 rejections, take a hard look at your technical and soft skills. What’s your education (degree & field)? In what areas have you struggled during your interviews?
Thanks. How would you define culture fit? is it an overall intuition or feel. We hear alot about team fit etc but what characteristics really define a fit? is it relevant experience of what the team is doing!
You forgot the most important part, SQL.
lol, just added!
maybe 20% of a data scientists job is related to building models. the rest has to do with how well you communicate, how you deal with ambiguity, how good you are at project management, working with senior leadership on product decision. building models is the easy part. spend more time emphasizing the hard stuff in interviews.
🍿