Data Science
& AI syllabus.

Understand how language models work, then run open-source models on your own hardware.

• Transformers, self-attention, embeddings and positional encoding
• Tokenization, pre-training and in-context learning
• Local inference with Llama, Gemma and GGUF
• Quantization and parameter-efficient adaptation with LoRA

Work through the complete machine learning lifecycle, from raw data to a model you can evaluate.

• Data cleaning and transformation with NumPy and Pandas
• Supervised learning: regression and classification
• Unsupervised learning and clustering
• Precision, recall, confusion matrices and neural network basics

Give AI models useful context and make their outputs consistent and measurable.

• Roles, constraints and few-shot prompting
• Reusable context packs and rules
• Structured JSON outputs
• Evaluation rubrics and guardrails

Connect language models to your own documents and data with retrieval-augmented generation.

• Document ingestion, semantic chunking and overlap
• Embeddings and vector search
• Retrieval across documentation and source data
• Building a knowledge retrieval system

Test AI systems against attacks and reduce risks before putting them into use.

• Prompt injection and adversarial testing
• Data leakage and context audits
• Guardrails and semantic security filters
• Documenting and evaluating defenses

Adapt a model to a specific domain and compare the results with retrieval-based approaches.

• Choosing fine-tuning or RAG
• LoRA and QLoRA
• Preparing domain-specific training data
• Tracking loss and evaluation metrics

Coordinate specialist agents and connect them to tools and external data.

• Agent orchestration patterns
• Shared state and memory
• Streaming agent and tool responses
• MCP servers for database and file access

Package, deploy and monitor AI systems so they can serve real users reliably.

• Reproducible deployments with Docker
• Cloud deployment and scaling
• Security, rate limiting and release checks
• Observability and model performance monitoring

Apply machine learning to a ride-hailing demand forecasting case study.

• Time-series forecasting
• Feature selection: location, timing and traffic
• Demand forecasts and dynamic pricing
• Measuring forecasts against actual outcomes

• CV and LinkedIn preparation
• A structured GitHub project portfolio
• Explaining your projects in interviews
• Technical and behavioral interview preparation