I applied online. The process took 4 weeks. I interviewed at GEICO (Washington, DC) in Aug 2015
Interview
I applied online and got a response within a day or two from Geico's HR. They scheduled a call with one of their data scientists. The call was about 30 mins long and went well. They asked me to talk about my data science/analytics experience. They explained that the next step would be a take-home test working on a hypothetical prediction model. I was given 4 days to work on this model and submitted my code and the results. This is where things went awry. I followed up a week later and HR told confirmed that they had received my submission but wouldn't say if they were still reviewing it or if they had chosen to not move ahead with my candidacy. Regardless of Geico's decision, I would've expected a proper response from an established company. After about a month, Geico did get back to me about continuing with the application process. However, by then I had already accepted a data scientist job at another company.
Interview questions [1]
Question 1
Describe one data science experience you've worked on.
I applied through a recruiter. I interviewed at GEICO in Oct 2022
Interview
HR was very patient. The whole process was smooth and fast. 4 rounds of interviews in total. The first round was ML+resume; the second was python + SQL + easy algorithm, the third round was real case machine learning models, the last round was with one of the managers and asked resume and one machine learning model.
Interview questions [1]
Question 1
Asked many machine learning questions such as decision tree, random forest, KNN, and Kmeans. Must know the detail of each model very well. Deep understanding of your resume, asked many details
Onsite technical screen had SQL and python coding questions. Format was in a double spaced google doc, which is by far the worst possible way to ask these questions. Interviewer didnt seem to understand the problems during the interview leading to confusion. GEICO, like many financial or insurance companies, have really poor interview processes in place for data science. Bottom quartile compensation = bottom quartile talent.
It was a very strenuous and long process which required a lot of time spending through multiple rounds of interviews and take home exam. Many technical questions regarding ML from multiple interviewers.