Data Science & ML with Gen AI Training

• Python fundamentals
• Functions
• OOP
• Exceptions
• Files
• APIs
• Virtual environments
• Git/GitHub

• Arrays
• DataFrames
• Cleaning
• Missing data
• Duplicates
• Transformation
• Aggregation
• Merging
• Feature engineering
  
  Project: Clean a messy real-world dataset.

• Probability
• Distributions
• Sampling
• Confidence intervals
• Hypothesis testing
• Correlation
• Regression
• A/B testing
   

• Matplotlib
• Seaborn
• Plotly
• Storytelling
• Interactive visualization
   

Supervised

• Linear Regression
• Logistic Regression
• Decision Trees
• Random Forest
• XGBoost
• SVM
• KNN

Unsupervised

• K-Means
• Hierarchical clustering
• PCA

• Train/validation/test
• Cross-validation
• Pipelines
• Feature selection
• Hyperparameter tuning
• Imbalanced data
• Model evaluation
• Explainability
• SHAP

• Text preprocessing
• TF-IDF
• Embeddings
• Sentiment analysis
• Classification
• NER
• Transformers
• BERT

Flagship

• LLM fundamentals
• Tokens
• Embeddings
• Prompt engineering
• Structured outputs
• Function calling
• RAG
• Vector databases
• LLM evaluation
• Hallucination
• Guardrails
  Tools: OpenAI API, Gemini, Claude, Hugging Face, LangChain, LlamaIndex, FAISS/Chroma

• Agent architecture
• Tool use
• Memory
• Planning
• Multi-step workflows
• Agent evaluation
• Human-in-the-loop

• MLflow
• FastAPI
• Docker
• Model serving
• CI/CD
• Monitoring

• Big Data architecture
• Spark
• PySpark
• Data pipelines
• Cloud data concepts

End-to-End AI Product

Example:

Nepal Tourism Demand Prediction + AI Travel Assistant

Data → EDA → ML → RAG → Agent → API → Docker → Deployment

Graduate must be able to:

Build, evaluate and deploy an ML/GenAI application rather than simply train a model in Jupyter.