Agentic AI, LLM engineering, and the models behind modern automation — multi-agent systems, LLM evaluation and comparisons, RAG, fine-tuning, AI infrastructure, security, and production AI engineering.
From the blog
Understand how LoRA and QLoRA enable fine-tuning of large language models on consumer hardware by training only a small fraction of parameters, with practical examples using Hugging Face and PEFT.
Implement the Map-Reduce pattern for AI agents to split large workloads across parallel agent workers and aggregate their results efficiently.
Learn how to build AI agents using the Mastra framework. This guide covers project setup, agent definition with typed tools, persistent memory, workflow orchestration, and deployment strategies for TypeScript-first agent applications.
Implement robust security for MCP servers with OAuth2 integration, API key validation, permission scopes, and token management to ensure AI agents only access tools they are authorized to use.
Build and deploy MCP servers accessible over HTTP using the Streamable HTTP transport, with Server-Sent Events for streaming, stateless vs stateful session modes, and strategies for horizontal scaling.
Master MCP prompts — server-defined instruction templates that guide AI agent behavior with dynamic arguments, context-aware instructions, and reusable prompt patterns across multiple agents.
Explore the Model Context Protocol specification from the ground up — JSON-RPC 2.0 message format, request/response lifecycle, transport layers, and how every MCP message is structured under the hood.
Learn how MCP resources differ from tools, how to define resource URIs and templates, expose read-only data to AI agents with proper content types, and implement pagination for large datasets.
Learn how MCP clients discover available servers, understand server manifests and tool catalogs, and build a lightweight server registry that lets agents dynamically connect to the right tools.
