AI & ML Topics

Quick Nav

---

Topics Checklist

LLM & GenAI Fundamentals

• Transformer architecture (attention, embeddings)
• Prompt engineering — zero-shot, few-shot, chain-of-thought
• RAG (Retrieval-Augmented Generation)
• Fine-tuning vs RAG vs prompt engineering trade-offs
• Evaluation — BLEU, ROUGE, LLM-as-judge
• Hallucination causes and mitigations

LangChain

• LCEL (LangChain Expression Language)
• Chains — basic, sequential, router
• Agents & tools
• Memory (ConversationBufferMemory, VectorStoreRetriever)
• Vector stores — FAISS, Chroma, Pinecone
• Document loaders, text splitters
• LangGraph — stateful multi-agent workflows
• LangSmith — tracing & evaluation

MCP (Model Context Protocol)

• What MCP is — tools, resources, prompts
• Architecture — host / client / server
• LangChain MCP integration
• Common MCP servers (filesystem, GitHub, DB, etc.)

ML Fundamentals (for interviews / AWS exam)

• Supervised vs unsupervised vs reinforcement learning
• Bias-variance trade-off
• Regularisation — L1 (Lasso), L2 (Ridge)
• Cross-validation
• Classification metrics — precision, recall, F1, AUC-ROC
• Common algorithms — linear regression, logistic regression, decision tree, random forest, gradient boosting, SVM, k-means
• Feature engineering, normalisation, handling missing values

MLOps

• Model lifecycle — training, evaluation, deployment, monitoring
• Feature stores
• Model versioning & registry
• Data drift vs model drift
• A/B testing models
• CI/CD for ML pipelines

System Design Notes (AI-Specific)

• [[RAG & LLM System]] — RAG architecture, chunking, retrieval, reranking
• [[ML Feature Store]] — offline vs online store, point-in-time correctness