Big Tech Career Transition

Create and activate a Python virtual environment and run a script in VS Code on macOS.

Use Python variables, primitive data types, arithmetic operators, comparison operators, and Boolean operators.

Transform and inspect text using Python string operations and methods.

Control program execution using conditional statements.

Process repeated operations using for loops and while loops.

Store, retrieve, and update ordered data using lists and tuples.

Store, retrieve, and update unique or keyed data using sets and dictionaries.

Define and call functions with parameters, return values, and local variables.

Import and use standard-library modules and third-party Python packages.

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

Interpret Python tracebacks and handle expected failures with exceptions.

Define and instantiate Python classes with attributes and methods.

Navigate files and run programs using essential macOS terminal commands.

Track and restore code changes using a local Git repository and commits.

Estimate the time complexity of basic list, dictionary, sorting, and searching operations.

Write and run unit tests for Python functions with pytest.

Manipulate equations, functions, powers, exponentials, and logarithms used in machine learning formulas.

Compute vector addition, scalar multiplication, dot products, and vector norms.

Compute matrix products and transposes and identify compatible matrix dimensions.

Calculate derivatives, partial derivatives, and gradients of simple functions.

Compute and interpret mean, median, variance, standard deviation, covariance, and correlation.

Calculate conditional probabilities and apply Bayes’ rule to simple events.

Model simple events using Bernoulli, binomial, and normal probability distributions.

Apply gradient descent to minimize a one-variable or two-variable loss function.

Create, reshape, index, and aggregate NumPy arrays using vectorized operations.

Load, select, filter, sort, and aggregate tabular data with pandas.

Correct missing values, duplicate records, incorrect types, and inconsistent categories in a tabular dataset.

Visualize distributions and relationships using Matplotlib and Seaborn.

Retrieve and filter relational data using SELECT, WHERE, ORDER BY, and LIMIT.

Combine related tables using SQL joins.

Summarize grouped data using SQL aggregate functions and GROUP BY.

Translate a business question into a prediction target, input features, and a measurable success criterion.

Create training, validation, and test partitions while preventing data leakage.

Train and evaluate a dummy model as a performance baseline.

Fit a linear regression model and interpret its coefficients and residuals.

Control linear-model complexity using L1 or L2 regularization.

Fit a logistic regression model and convert predicted probabilities into class labels.

Train and interpret a decision-tree model for classification or regression.

Train a random-forest model and compare it with a single decision tree.

Train a gradient-boosted tree model using a scikit-learn-compatible library.

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

Evaluate classification predictions using a confusion matrix, precision, recall, and F1 score.

Select a classification threshold using ROC and precision-recall curves.

Estimate generalization performance using k-fold cross-validation.

Tune model hyperparameters using grid search or randomized search.

Impute missing values, encode categories, and scale numeric features with a ColumnTransformer.

Combine preprocessing and estimation steps in a reusable scikit-learn Pipeline.

Measure global feature importance using permutation importance.

Diagnose underfitting and overfitting using training and validation learning curves.

Address imbalanced classification using class weights and appropriate evaluation metrics.

Perform slice-based error analysis to identify weak model behavior on subgroups.

Explain individual predictions using SHAP values.

Group unlabeled observations using k-means clustering and assess cluster quality.

Reduce feature dimensions using principal component analysis while tracking explained variance.

Build a baseline text classifier using TF-IDF features and a linear model.

Evaluate a time-series forecasting baseline using chronological validation and lag features.

Create and manipulate PyTorch tensors with explicit shapes, data types, and devices.

Compute gradients for tensor operations using automatic differentiation.

Construct a multilayer perceptron with suitable activation and output layers.

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

Reduce neural-network overfitting using dropout, weight decay, or early stopping.

Diagnose neural-network training using loss curves and validation metrics.

Train a convolutional neural network for image classification.

Adapt a pretrained neural network to a new classification task using transfer learning.

Define a tabular data contract containing field types, allowed values, and quality rules.

Build an idempotent extract-transform-load pipeline in Python.

Store and query partitioned tabular data in Parquet format.

Automate schema, range, uniqueness, and missing-value checks in a data pipeline.

Transform and aggregate a dataset using the PySpark DataFrame API.

Write SQL window functions for ranking and rolling calculations.

Orchestrate dependent pipeline tasks with scheduling, retries, and failure handling.

Choose between batch and streaming processing based on freshness, scale, and complexity requirements.

Organize training, inference, configuration, and test code as an installable Python project.

Track model parameters, metrics, and artifacts using MLflow experiments.

Register and version an approved model artifact in a model registry.

Expose model predictions through a validated FastAPI endpoint.

Package an inference service as a reproducible Docker image.

Apply a branch-and-pull-request workflow in a shared GitHub repository.

Run tests and build a container automatically in a continuous-integration workflow.

Monitor a deployed model for service errors, latency changes, data drift, and prediction-quality degradation.

Map storage, compute, container, identity, training, registry, endpoint, and monitoring components across Azure, Google Cloud, and AWS.

Configure least-privilege access and budget safeguards for a machine-learning sandbox account.

Submit a managed training job and register its model in Azure Machine Learning.

Deploy and test a managed online endpoint in Azure Machine Learning.

Run a managed training pipeline and register its model in Google Vertex AI.

Deploy and test an online prediction endpoint in Google Vertex AI.

Compare Azure and Google Cloud deployment choices by latency, scalability, operational effort, and cost.

Trace an equivalent machine-learning lifecycle through Amazon S3, ECR, SageMaker training, Model Registry, and endpoints.

Write a capstone design brief with a user need, data plan, success metric, constraints, and deployment target.

Produce a model-selection report that compares experiments and justifies the final capstone model.

Publish a reproducible GitHub case study with setup instructions, an architecture diagram, a model card, and a working demonstration.

Tailor résumé bullets to demonstrate measurable machine-learning, software-engineering, cloud, and transferable customer-service skills.

Solve entry-level Python interview problems using arrays, strings, hash maps, and explicit complexity analysis.

Solve machine-learning-oriented SQL interview problems using joins, aggregations, subqueries, and window functions.

Design a machine-learning system from requirements covering data, training, serving, scaling, and monitoring.

Answer behavioral interview questions with concise STAR stories drawn from call-centre work and course projects.