Daily Dose of Data Science | Avi Chawla | Substack·1y9 RAG, LLM, and AI Agent Cheat Sheets
This post provides visual cheat sheets for AI engineers covering various topics, including Transformer vs. Mixture of Experts in LLMs, fine-tuning techniques, RAG vs Agentic RAG, strategies for chunking in RAG, levels of agentic AI systems, and more. These resources are designed to help cultivate essential skills for developing impactful AI and ML systems in the industry.
Machine Learning News·1yOpenAI Releases a Practical Guide to Building LLM Agents for Real-World Applications
OpenAI has released a guide for engineering and product teams on building autonomous AI systems. The guide describes the essential components of an AI agent, appropriate use cases, and technical foundations using the OpenAI Agents SDK. It also emphasizes safety and human oversight mechanisms to ensure reliable and controllable deployment of AI agents.
Medium·1yBuilding Multi-Agent with Google's A2A (Agent2Agent) Protocol, Agent Development Kit(ADK), and MCP (Model Context Protocol) - A Deep Dive(Full Code)
Explore building a multi-agent AI application using Google’s A2A protocol, ADK, and MCP. Delve into core concepts like task-based communication, agent discovery, framework-agnostic interoperability, multi-modal messaging, and standardized message structures. Learn about a practical demo involving a multi-agent travel planner that showcases these principles in action.
Daily Dose of Data Science | Avi Chawla | Substack·1yAI Agents Crash Course—Part 8 and 9
Parts 8 and 9 of the AI Agents crash course focus on the significance of Memory in AI agents. It covers how memory enables context-awareness, various types of memory (Short-Term, Long-Term, Entity, Contextual, User), and their unique purposes. The posts provide theoretical, practical insights, and implementation details for each memory type.
Daily Dose of Data Science | Avi Chawla | Substack·1yImplement Planning Agentic Pattern from Scratch
Part 11 of the AI Agents crash course introduces the Planning agentic pattern, focusing on implementing it from scratch using Python and a Language Learning Model (LLM). The post highlights the importance of structured planning to improve the thoroughness of LLM decisions, covering recent research, the Planning loop pattern, and best practices. It also includes detailed instructions for creating a system prompt and a lightweight agent class, along with examples of both manual and automated Planning loops.
Daily Dose of Data Science | Avi Chawla | Substack·1yGuardrails for AI Agents
The post explains how reinforcement fine-tuning (RFT) enhances open-source LLMs, offering accuracy gains and efficient fine-tuning with few examples. It also details implementing guardrails for AI agents to prevent issues like hallucination and infinite loops. The guide walks through setting up validation checkpoints, limiting tool usage, and specifying fallback mechanisms with practical code examples.