Prompting for Effective LLM Reasoning with Gemini

Overzicht

This course equips you with strategies to harness the power of Google Cloud’s Gemini for agentic AI development. You will explore the fundamentals of agentic AI and advanced prompting techniques such as role-based prompting and chain-of-thought (CoT) prompting.

The course covers prompt instruction refinement and chaining methods for enhancing reasoning capabilities. Additionally, you will learn to implement feedback loops specifically for code generation tasks.

For the final project, you will design a Legal Intelligence AI System, applying learned techniques to solve real-world legal challenges.

Lesprogramma

• Introduction to Agentic AI with Gemini and Google Cloud Platform
Explore agentic AI concepts using Gemini models and Google Cloud Platform, covering prerequisites, environment setup, and key features for effective LLM reasoning.
• Role-Based Prompting
Explains the theory of using roles or personas to control the tone, style, and expertise of an LLM's output.
• Implementing Role-Based Prompting with Vertex AI Gemini
Learn to create expert business personas for Vertex AI Gemini, guiding the model with role-based prompts to produce domain-specific, high-quality, specialized AI responses.
• Chain-of-Thought and ReACT Prompting
Explains the conceptual frameworks for Chain-of-Thought (CoT) for guided reasoning and ReAct (Reason+Act) for enabling agents to plan and take actions.
• Implementing Chain-of-Thought and ReACT Prompting with Vertex AI Gemini
Learn to build iterative, reasoning agents with Chain-of-Thought and ReACT prompting using Vertex AI Gemini to solve complex, multi-step business problems systematically.
• Prompt Instruction Refinement
Explains the theory of systematically refining prompt instructions by modifying components like Role, Task, Context, Examples, and Output Format.
• Implementing Prompt Refinement with Vertex AI Gemini
Learn to refine and optimize AI prompts using Vertex AI Gemini, applying systematic quality metrics like clarity, specificity, completeness, and structure.
• Chaining Prompts for Agentic Reasoning
Explains the conceptual framework for building multi-step AI workflows by linking the output of one prompt to the input of the next, and the importance of validation.
• Implementing Prompt Chaining with Vertex AI Gemini
Learn to implement prompt chaining with Vertex AI Gemini for agentic reasoning: build sequential workflows, manage context, handle errors, and ensure quality in multi-step tasks.
• LLM Feedback Loops
Explains the conceptual framework for building self-improving systems where an agent uses feedback from its own actions to iteratively refine its output.
• Implementing LLM Feedback Loops with Vertex AI Gemini
Learn to implement automated feedback loops for code generation using Vertex AI Gemini, covering self-validation, intelligent retries, quality gates, and production monitoring.
• Project: Legal Intelligence AI System
Learn prompting techniques to build an agent that can produce consistent business intelligence reports with built-in quality checks and transparent reasoning traces that analysts can review.

---

Gegeven door

Brian Cruz and Noble Ackerson

---

Vakgebieden

Artificial Intelligence