BSDA4001 - Data Science and AI Lab

This is a hands-on course designed to bridge the gap between academic knowledge and industry needs in Data Science and AI. The course revisits essential data science concepts and includes contemporary AI technologies. Each week introduces a new concept and accompanies an end-to-end assignment, simulating real-world workflows.

Key tools: NumPy, Pandas, Matplotlib, Seaborn, Jupyter, Git | | Week 2 | Week 2: Machine Learning with Scikit-learn

Core models: regression, classification, clustering

Pipelines, hyperparameter tuning, evaluation metrics | | Week 3 | Week 3: Deep Learning with PyTorch and TensorFlow

Building and training deep neural networks

Model saving/loading, GPU training | | Week 4 | Week 4: Computer Vision and Image Processing

Using OpenCV, PIL, PyTorch/TensorFlow for image loading, preprocessing

Transfer learning with CNNs (Re | | Week 5 | Week 5: NLP and LLMs using HuggingFace

Using BERT, RoBERTa for classification/generation

Tokenization, fine-tuning, and inference | | Week 6 | Week 6: Prompt Engineering, LangChain & LLM APIs

Working with LangChain, templates, and memory modules

Prompt tuning best practices

Using OpenAI, | | Week 7 | Week 7: Retrieval-Augmented Generation (RAG) & Vector DBs

Concepts of semantic search & embedding spaces

Use of FAISS, Pinecone, ChromaDB | | Week 8 | Week 8: Agents and Autonomous AI (Langchain, LangGraph)

Concepts of tool use, planning, multi-step task agents

Implementing basic agents using Lang | | Week 9 | Week 9: Data Visualization & Dashboards

Tools: PowerBI, Tableau, Streamlit, Plotly

Dashboard design, storytelling with data | | Week 10 | Week 10: Time Series Forecasting

Exploratory Analysis of Time Series

Forecasting using Facebook Prophet and DeepAR Trend, seasonality, holidays, ano | | Week 11 | Week 11: Big Data and Cloud Platforms - Google Cloud Platform

Build and run an end-to-end pipeline in GCP

Working with GCP, BigQuery, Vertex AI

Con | | Week 12 | Week 12: Model Explainability, Ethics & Mini-project

Tools: SHAP, LIME for understanding predictions

Concepts: data privacy, fairness, AI ethics Min |