We help companies handpick the best talent for their specific job profiles. Depending on the specific roles and responsibilities of the job, candidates go through a rigorous multi-stage adaptive testing process, including taking our proprietary exams and interviewing with our experts, where they are evaluated based on a multitude of factors. For example, we outline here some of the criteria we recently took into account to shortlist a mid-level data scientist for a job requiring facility with text mining, recommender systems, and time analysis.
- understanding of deep learning methods and level of comfort dealing with intricacies of modern deep learning architectures
- ability and preparedness to code and implement, e.g., in Python as well as deep learning libraries and wrappers such as PyTorch/Tensorflow/Keras
- aptitude to critically analyze the pros and cons (e.g., computational, statistical issues etc.) of different machine learning/deep learning/NLP/temporal (time-series) algorithms in specific scenarios or tasks
- facility in designing an entire pipeline (e.g., for a real world problem such as designing their own version of Amazon/eBay recommendation system or Google translate system):
- defining and formulating real world problems mathematically as machine learning/optimization/statistical objectives
- preprocessing/analyzing data for insights
- design appropriate machine learning/NLP/deep learning models
- estimate uncertainty in predictions
- conduct statistical hypothesis tests for deployment
- ability to stay calm and think through seemingly difficult/tricky challenges and problems
- identifying key factors that could determine their decision if they are extended an offer