AI & Data Science Full-stack

Python basics, data types, loops and functions; working with files, CSV and JSON; virtual environments, Jupyter and Git.

• Python basics: variables, data types, conditions and loops
• Functions, modules and writing clean, readable code
• Lists, dictionaries and working with real records
• Reading and writing files: CSV, JSON and Excel
• Setting up: virtual environments, Jupyter and VS Code
• Version control with Git and GitHub

pandas DataFrames, cleaning and joining data, group-by and pivot tables; SQL queries, joins and window functions.

• pandas DataFrames: selecting, filtering and sorting data
• Cleaning data: missing values, duplicates and wrong types
• Joining tables, group-by summaries and pivot tables
• SQL basics: SELECT, WHERE, GROUP BY and ORDER BY
• SQL joins, subqueries and window functions
• Answering business questions with data and presenting them clearly

Descriptive statistics, probability, sampling, hypothesis tests and A/B testing; charts with Matplotlib, Seaborn and Plotly.

• Descriptive statistics: averages, spread and distributions
• Probability and sampling, and why sample size matters
• Hypothesis tests, confidence intervals and p-values
• A/B testing for product and marketing decisions
• Charts with Matplotlib and Seaborn
• Interactive dashboards with Plotly

Supervised learning, regression and classification, train/test splits and cross-validation, scikit-learn pipelines.

• What machine learning is, and when not to use it
• Regression: predicting numbers
• Classification: predicting categories
• Train/test splits, cross-validation and overfitting
• scikit-learn pipelines from raw data to predictions
• Comparing models and explaining what drives a prediction

Feature engineering, imbalanced data, ensembles (random forests, XGBoost, LightGBM), metrics, explainability with SHAP.

• Creating useful features from dates, text and categories
• Handling imbalanced data
• Ensembles: random forests, XGBoost and LightGBM
• Choosing the right metric: precision, recall, AUC and cost
• Explaining models with SHAP
• Fairness checks for decisions about people

Neural networks, training loops, CNNs for images, transfer learning, GPUs on Google Colab.

• How neural networks learn: layers, loss and gradient descent
• Writing training loops in PyTorch
• Convolutional networks for images
• Transfer learning with pre-trained models
• Training on free GPUs with Google Colab
• Avoiding overfitting: augmentation, dropout and early stopping

Text cleaning, embeddings, transformers and Hugging Face, fine-tuning for classification, multilingual models.

• Cleaning and preparing text
• Word embeddings and sentence embeddings
• Transformers and the Hugging Face ecosystem
• Fine-tuning a model for text classification
• Working with Hindi and other Indian languages
• Sentiment analysis and topic modelling

Trends and seasonality, ARIMA and Prophet, gradient-boosted forecasts, evaluating forecasts over time.

• Trends, seasonality and festival effects in time data
• Classic forecasting with ARIMA
• Forecasting with Prophet
• Gradient-boosted forecasts with lag features
• Back-testing: evaluating forecasts over time
• Comparing against simple baselines

How LLMs work, prompting, embeddings and vector databases, retrieval-augmented generation, evaluating answers.

• How large language models work, in plain terms
• Prompting techniques for reliable answers
• Embeddings and vector databases
• Retrieval-augmented generation (RAG) with sources
• Evaluating answers for accuracy and faithfulness
• Cost, privacy and safety when using LLMs

Function calling, multi-step agents, planning and memory, guardrails and evaluation of agent behaviour.

• Function calling: letting models use tools
• Multi-step agents that plan and act
• Memory and keeping track of context
• Agent frameworks such as LangGraph and CrewAI
• Guardrails and human approval steps
• Testing and evaluating agent behaviour

APIs with FastAPI, Docker, CI/CD, model registries, monitoring and data drift, cloud deployment basics.

• Serving models as APIs with FastAPI
• Packaging with Docker
• Automated tests and CI/CD pipelines
• Model registries and versioning
• Monitoring, data drift and retraining
• Cloud deployment basics

Scoping a real problem, project planning, portfolio, résumé and GitHub reviews, mock technical interviews.

• Choosing and scoping a real problem
• Planning a project with milestones
• Building a portfolio and GitHub profile
• Résumé and LinkedIn reviews
• Mock technical and HR interviews
• Presenting your capstone to a panel