generative-ai

Provides a curated, multi-stage learning progression from foundational AI/ML/DL concepts through transformer architectures, LLM fundamentals, prompt engineering, RAG systems, and agentic AI frameworks. The learning path is organized as interconnected modules with prerequisite dependencies, enabling learners to build mental models incrementally before tackling advanced implementations. Uses Jupyter Notebooks and markdown documentation to combine theory with executable code examples.

Implements Retrieval Augmented Generation systems that integrate document retrieval with LLM generation, including guidance for selecting appropriate embedding models based on use-case requirements (semantic similarity, multilingual support, domain-specific performance). The system evaluates RAG quality through metrics and supports multiple LLM providers (OpenAI, Anthropic, Ollama) and cloud platforms (AWS, Azure, Google VertexAI). Uses vector storage and semantic search to retrieve relevant context before generation.

Provides curated recommendations for GenAI technology stacks including LLM aggregators, agentic frameworks, AI coding assistants, and cloud integrations. Compares tools across dimensions like ease of use, feature completeness, community support, and cost. Helps teams select complementary tools that work well together rather than evaluating tools in isolation.

Provides implementations and comparison of agentic AI frameworks (CrewAI, LangGraph) that enable autonomous agents to decompose tasks, call tools, and iterate toward solutions. Includes patterns for agent design, tool integration, and multi-agent orchestration. Supports both simple sequential agents and complex reasoning chains with memory and state management across multiple steps.

Demonstrates a production-grade application integrating chat, OCR (optical character recognition), RAG, and agentic AI capabilities into a single Llama 4-based system. The app uses a modular architecture where each capability (chat, document processing, information retrieval, autonomous reasoning) can be invoked independently or composed together. Includes environment configuration, requirements management, and evaluation utilities for measuring system performance.

Provides deployment guides and implementation examples for deploying Generative AI solutions across AWS, Azure, and Google VertexAI platforms. Includes platform-specific patterns for model serving, API integration, authentication, and cost optimization. Abstracts platform differences to enable multi-cloud or cloud-agnostic deployments where possible.

Provides comprehensive prompt engineering guidance with executable examples using Ollama-based models and other LLM providers. Covers techniques like chain-of-thought prompting, few-shot learning, role-based prompting, and structured output formatting. Includes notebooks demonstrating how different prompt structures affect model behavior and output quality across different model families.

Provides a decision framework and comparison notebook for selecting appropriate embedding models based on use-case requirements (semantic similarity, multilingual support, domain-specific performance, latency, cost). Evaluates embedding models across dimensions like vector dimensionality, inference speed, and performance on domain-specific benchmarks. Includes code for measuring embedding quality and comparing models empirically.

Provides comprehensive interview preparation resources including agentic AI interview questions and general GenAI interview questions in PDF format. Covers conceptual understanding, implementation patterns, system design, and practical problem-solving across GenAI domains. Organized by topic and difficulty level to support preparation for various seniority levels.

Provides guidance and implementations for abstracting LLM provider differences (OpenAI, Anthropic, Ollama, Google VertexAI) behind a unified interface. Enables switching between providers without changing application code, supporting cost optimization, redundancy, and experimentation across different models. Includes documentation on provider-specific features, pricing, and performance characteristics.

Provides structured guidance for the complete lifecycle of GenAI projects from ideation through deployment and monitoring. Covers project planning, technology stack selection, implementation patterns, testing strategies, and production deployment considerations. Organized as best practices and guidelines rather than prescriptive rules, enabling teams to adapt patterns to their specific context.