AI

Configure a Python environment and run code in scripts and Jupyter notebooks.

Use Python variables, scalar data types, expressions, and type conversions.

Write conditional statements to implement decision logic.

Write for and while loops to process repeated operations.

Select and manipulate lists, tuples, dictionaries, and sets for a given task.

Define and call functions with parameters, return values, and clear docstrings.

Use comprehensions to construct transformed and filtered collections.

Read and write text, CSV, and JSON data with Python.

Navigate directories, manage files, and execute Python programs from a command line.

Create virtual environments and install pinned Python packages.

Organize reusable Python code with imports and modules.

Define a Python class with attributes and methods for a simple data-processing component.

Diagnose Python failures by interpreting tracebacks and inspecting program state.

Write unit tests for Python functions with pytest.

Perform indexed and vectorized calculations on NumPy arrays.

Track and share code changes using a Git and GitHub commit workflow.

Represent data as vectors and perform vector addition and scalar multiplication.

Compute and interpret dot products between vectors.

Calculate vector norms and distances used by machine-learning algorithms.

Interpret a matrix as both a data table and a linear transformation.

Perform matrix multiplication and verify dimensional compatibility.

Solve a system of linear equations using Gaussian elimination.

Determine the rank and linear independence of a set of vectors.

Compute and interpret eigenvalues and eigenvectors of a simple matrix.

Evaluate, compose, and graph single-variable functions used in machine learning.

Interpret a derivative as a local rate of change and tangent slope.

Differentiate functions using power, product, quotient, and exponential rules.

Compute partial derivatives of multivariable functions.

Compute and interpret the gradient of a scalar-valued function.

Apply the chain rule through a simple computational graph.

Identify local minima and maxima using first- and second-derivative information.

Perform gradient-descent updates and analyze the effect of the learning rate.

Calculate joint, marginal, and conditional probabilities from events or tables.

Apply Bayes’ rule to update a probability after observing evidence.

Model count and binary outcomes with common discrete probability distributions.

Model continuous measurements with uniform, normal, and exponential distributions.

Compute and interpret the expectation and variance of a random variable.

Compute and distinguish covariance and correlation between variables.

Construct and interpret a confidence interval for a population mean.

Estimate model parameters using maximum likelihood.

Identify tables, rows, columns, primary keys, and foreign keys in a relational schema.

Retrieve, filter, and sort records with SQL SELECT queries.

Summarize grouped data with SQL aggregate functions.

Combine related tables using SQL joins.

Load and inspect tabular data with pandas DataFrames.

Correct data types and handle missing or duplicated observations with pandas.

Transform columns and calculate grouped summaries with pandas.

Visualize distributions and relationships to identify data quality issues and patterns.

Translate a prediction problem into features, a target, and train-validation-test partitions.

Fit a linear regression model by minimizing squared error.

Evaluate regression predictions using residuals, MAE, RMSE, and R-squared.

Fit regularized linear models and interpret the effect of regularization strength.

Fit a logistic regression model and interpret its predicted probabilities.

Evaluate classifiers with confusion matrices, precision, recall, F1, and ROC-AUC.

Encode categorical features and scale numeric features with scikit-learn transformers.

Construct a leakage-safe scikit-learn pipeline for preprocessing and model fitting.

Fit and interpret a decision tree for a supervised-learning task.

Control decision-tree overfitting with depth, leaf-size, and pruning settings.

Fit a random forest and explain how bagging reduces model variance.

Fit a gradient-boosted tree model and explain sequential error correction.

Compare candidate models using k-fold cross-validation.

Tune model hyperparameters with grid or randomized search.

Diagnose underfitting and overfitting with training and validation learning curves.

Select a classification threshold from a precision-recall trade-off.

Create domain-informed numeric, categorical, and interaction features without target leakage.

Measure model reliance on input features with permutation importance.

Cluster standardized observations with k-means.

Evaluate cluster separation with silhouette scores.

Reduce feature dimensionality with principal component analysis.

Detect unusual observations with an isolation forest.

Create time-ordered validation splits that prevent look-ahead bias.

Train and evaluate a lag-feature-based forecasting model.

Create and manipulate PyTorch tensors on CPU and GPU devices.

Compute the forward pass of a dense neural-network layer.

Compare activation functions and select one for a hidden or output layer.

Construct a multilayer perceptron for a supervised-learning task.

Compute gradients through a neural network using backpropagation and automatic differentiation.

Load mini-batches with PyTorch Dataset and DataLoader components.

Implement a PyTorch training loop with a loss function and optimizer.

Use validation loss and early stopping to control neural-network overfitting.

Structure an ML repository with separate configuration, training, inference, and test components.

Serialize and reload a fitted preprocessing-and-model pipeline for deterministic inference.

Write unit and integration tests for data validation and model inference.

Track experiment parameters, metrics, and artifacts with MLflow.

Expose model predictions through a validated FastAPI endpoint.

Package an inference service as a Docker container.

Run automated tests and container builds with a GitHub Actions workflow.

Deploy a containerized inference service through Amazon ECR and AWS App Runner.

Translate a predictive-maintenance need into a prediction target, unit of prediction, and success metric.

Audit equipment sensor data for schema violations, leakage, and unsuitable observations.

Establish a reproducible baseline for the predictive-maintenance task.

Compare tracked candidate models and select one using predictive and operational criteria.

Perform slice-based error analysis across equipment types and operating conditions.

Deploy the selected model by applying the tested AWS inference workflow.

Define service-health, data-drift, and retraining checks for the deployed model.

Present the project’s architecture, evidence, limitations, and trade-offs in an interview-style portfolio walkthrough.