AI Agent Development with Python

Explore the agent execution loop, where observe gathers input and context, think analyzes, decide selects actions, act uses tools, and respond loops toward the goal.

Explore how prompt engineering shapes ai agent behavior through system prompts, personas, tone, and constraints, then design role, objective, rules, and output formats for reliable interactions.

Connect Python to a large language model via an API, send user messages, and preserve conversation history to maintain context in a simple chat loop.

Design reliable agents by enforcing structured outputs and a clear schema, then validate responses, handle edge cases, and apply guardrails with retries and fallbacks.

Demonstrates building a local personal ai assistant with python, olama, streamlet, pydantic, pymu pdf, and sqlite; shows setup, virtual env, running the app, and chat tasks.

Explore AI agent planning architectures, including the planner-executor and react-reasoning patterns, which break goals into steps, reason, use tools, and adapt. Learn when each pattern fits complex or dynamic tasks.

Explore reflection, routing, and decision logic to build ai agents that learn from feedback, self-correct outputs, and route tasks to the right tools for reliable performance.

Design safe and reliable planning for ai agents by defining goals within guardrails, specifying autonomy levels, validating outputs, and monitoring decisions to ensure safety and appropriate autonomy.

Participate in a hands-on lab to develop an ai task planning agent using Python, applying practical task planning concepts and agent design.

Demonstrates setting up and running the ai task planning agent, including creating a virtual environment, running the streamlit app, and generating, revising, and tracking a json plan.

Understand how AI agents move from conversation to action by using tools and function calls to perform real-world tasks. Explore tool schemas, validation, and safe routing.

Engage in a hands-on lab to build tool-using productivity agents with python, advancing ai agent development.

Explore how a tool-using productivity agent leverages calculator, weather, web search, and data reading tools within a streamlit app to run experiments and save results.

Learn how memory makes AI agents context aware by using short-term, long-term, episodic, semantic, and procedural memory. See how these types work together to transform one-time responses into context-aware behavior.

Explore persistent memory for AI agents using sqlite to save session-based conversations, roles, content, and timestamps, enabling coherent cross-session workflows with privacy controls and selective memory.

Learn how retrieval augmented generation (RAG) blends retrieval and generation to ground answers in trusted sources, delivering grounded, cited, and up-to-date responses from private knowledge.

Develop an ai agent with python through a hands-on lab that builds a memory-enabled document assistant.

Demonstrate memory-enabled document assistant development by uploading and ingesting pdfs with rag plus Chroma DB, querying documents via chat, and managing long-term memory with a memory level.

Master autonomous research by guiding a question through planning, searching, extraction, evaluation, reflection, and writing to produce a structured, evidence-based report with tracked sources.

An ai research agent uses reliability scoring: relevance, credibility, freshness, evidence strength, and consistency, to support fact-checking, duplicate detection, and a verification workflow for a trusted report.

The lecture presents iterative research loops that set a goal, search for evidence, reflect on gaps with gap analysis, refine queries, and stop when evidence is credible and complete.

Build an autonomous research agent in a hands-on lab using Python, and apply core AI agent development concepts for practical experimentation.

Demonstrate an autonomous research agent that fetches sources and generates a final report with executive summary. Explore the environment setup and the full workflow trace.

Explore multi-agent fundamentals by comparing single-agent and multi-agent designs, and learn how planners, researchers, tool users, writers, reviewers, and a supervisor collaborate with a shared state to coordinate task handoffs.

Explore four core patterns in multi-agent systems: planner, executor, reviewer, and supervisor. Understand how these agents share state, organize goals into tasks, and ensure quality through evaluation and coordination.

Discover how to coordinate a team of specialized AI agents using shared state, structured messaging, and deliberate handoffs to improve reliability, reduce duplication, and preserve context.

Explore LangGraph workflow orchestration by turning separate agent steps into a structured, controllable graph with shared state, nodes, edges, and checkpoints for sequential or parallel execution.

Learn to build and orchestrate multiple AI agents with a hands-on lab focused on multi-agent content creation and analysis workflows using Python.

Explore how end-to-end agent architecture coordinates planning, tools, memory, specialized agents, and review to transform a user goal into a structured, validated deliverable.

Coordinate an ai workflow from user goal to final output with a workflow hub, state store, and shared state, keeping execution visible and controlled through human approvals for high-risk steps.

Learn how to build reliable, fast AI agent systems using logging, observability, and caching to optimize latency, cost, and accuracy. Measure, trace, and alert to continuously improve performance and reliability.

Build robust ai systems with guardrails, permissions, retries, and error recovery. Enforce input filtering, policy rules, action limits, output validation, escalation, and observability to ensure safe, auditable operations.

Engage in a hands-on lab to build an autonomous AI operations assistant using Python, applying agent development skills to create practical AI agents.

Demonstrates building and running an autonomous AI operations assistant with Python, including environment setup, private memory use, plan execution, and web search and memory tools to generate a final report.