Where can you find qualified {role_name}?

What are the best practices for headhunting {role_name}?

How to shortlist candidates?

Once you have started to get applications from applicants, a thorough screening process and shortlisting of prospects will help you make the most of your time spent with the most qualified ones. 

Automated shortlisting tools :

Automated screening quickly filters out unqualified candidates, saving time for manual review. This allows the manual process to focus on the most promising candidates, ensuring the best ones are considered for further evaluation.

Screening questions to auto-shortlist based on predefined criteria

like qualifications, location, experience, and skills. Either use job board or use an ATS such as whitecarrot. Here are some questions for {role_name}

• How many years of experience do you have with machine learning algorithms?
  • Auto-reject: Less than 2 years.
• How many years of experience do you have using Python for machine learning?
  • Auto-reject: Less than 1 year.
• Are you proficient with machine learning frameworks like TensorFlow or PyTorch?
  • Auto-reject: No experience with these frameworks.
• Do you have experience deploying machine learning models in production?
  • Auto-reject: No production deployment experience.
• Are you located within [specified location] or willing to work remotely?
  • Auto-reject: Not willing to relocate or work remotely.

Skill based question to auto shortlist candidate

Analyze the skill test data to automatically shortlist top-performing applicants. (recommended screening test time - 15 minutes). Here are some skill test questions for {role_name}

Machine Learning Algorithms

Test understanding of common machine learning algorithms.

• Q1: Which of the following is a supervised learning algorithm?
  • a) K-Means Clustering
  • b) Decision Trees
  • c) Principal Component Analysis
  • d) DBSCAN
  • Correct Answer: b) Decision Trees
• Q2: What is the purpose of cross-validation in machine learning?
  • a) Reducing model complexity
  • b) Evaluating the performance of a model
  • c) Selecting hyperparameters
  • d) Training the model with more data
  • Correct Answer: b) Evaluating the performance of a model
• Q3: Which evaluation metric is most suitable for imbalanced classification problems?
  • a) Accuracy
  • b) Precision
  • c) F1 Score
  • d) Mean Squared Error
  • Correct Answer: c) F1 Score

Python

Test knowledge of Python for data science and machine learning.

• Q1: Which library is commonly used for data manipulation in Python?
  • a) Pandas
  • b) NumPy
  • c) Matplotlib
  • d) TensorFlow
  • Correct Answer: a) Pandas
• Q2: What does the function train_test_split from scikit-learn do?
  • a) Splits data into training and testing datasets
  • b) Normalizes the dataset
  • c) Performs cross-validation
  • d) Trains the model
  • Correct Answer: a) Splits data into training and testing datasets
• Q3: How do you calculate the accuracy of a model in Python using scikit-learn?
  • a) accuracy(y_test, y_pred)
  • b) metrics.accuracy_score(y_true, y_pred)
  • c) evaluate.accuracy(y_true, y_pred)
  • d) metrics.mean_squared_error(y_true, y_pred)
  • Correct Answer: b) metrics.accuracy_score(y_true, y_pred)

Model Deployment

Evaluate the candidate’s experience with deploying models.

• Q1: Which tool can be used to deploy machine learning models as RESTful APIs?
  • a) Flask
  • b) Pandas
  • c) Jupyter
  • d) Seaborn
  • Correct Answer: a) Flask
• Q2: What is a key advantage of using Docker for model deployment?
  • a) Ensures reproducibility and isolation of the model environment
  • b) Improves model accuracy
  • c) Reduces data preprocessing time
  • d) Increases the training speed of the model
  • Correct Answer: a) Ensures reproducibility and isolation of the model environment
• Q3: How can you deploy a machine learning model on AWS using SageMaker?
  • a) Write the model using SQL
  • b) Upload the model to SageMaker and create an endpoint for deployment
  • c) Use Pandas to import the model
  • d) Create a Spark job to deploy the model
  • Correct Answer: b) Upload the model to SageMaker and create an endpoint for deployment

Note - Auto reject candidates if scores less than 70% in this section

One way video interview

Use tools like hirevue, whitecarrot.io to ask candidates pre-recorded questions about their experience and skills.

Use sample question given in scorecard.

Collect other information 

Collect data from shortlisted candidates, such as salary expectations and visa status.

Manual candidate profile shortlisting:

Thoroughly review the CVs of the top scoring candidates from the automated process

Look for evidence of the required skills, experience, and achievements

Review the candidate’s portfolio or GitHub repositories to see examples of their work.