website: metadata: name: "The AI PM Builder - Syllabus - Google Reichman AI Tech School" title: "The AI PM Builder: Prototype & Validate, Evaluate & Ship" core_info: source_url: "https://graitechschool.co.il/the-ai-pm-builder-syllabus" organization: "Google Reichman AI Tech School" note: "changes may apply" description: "As a product team member, either a PM, PD, Developer or Data analyst, your reality has changed. Learn the essential tools & processes of AI PMs that will empower you to contribute more to your team and lead with AI. In these two modules, we will use Claude Code and dive into automated discovery, AI PRDs, design & prototyping, evals & guardrails, agent architectures, and agentic unit economics, together forming the core skills needed by product leaders in the design, prototyping, and deployment of AI products." header: primary_elements: [] navigation: [] actions: [] content: - section: title: "The AI PM Builder: Prototype & Validate, Evaluate & Ship" description: "As a product team member, either a PM, PD, Developer or Data analyst, your reality has changed. Learn the essential tools & processes of AI PMs that will empower you to contribute more to your team and lead with AI. In these two modules, we will use Claude Code and dive into automated discovery, AI PRDs, design & prototyping, evals & guardrails, agent architectures, and agentic unit economics, together forming the core skills needed by product leaders in the design, prototyping, and deployment of AI products." elements: - type: "image" alt: "Image 1" url: "https://www.graitechschool.co.il/wp-content/uploads/2023/02/Group-11870.png" - type: "image" alt: "Image 2" url: "https://www.graitechschool.co.il/wp-content/uploads/2022/12/%D7%A0%D7%99%D7%A0%D7%94-%D7%A2%D7%A0%D7%AA-%D7%95%D7%90%D7%93%D7%99%D7%A1%D7%9C%D7%90%D7%9D-scaled.jpg" actions: [] - section: title: "Program Format" description: null elements: - type: "list" items: - "2 independent, hands-on modules (Prototype & Validate, Evaluate & Ship)" - "4 sessions per course | 4 hours per session | Physical learning at Google Campus" - "Weekly practice session | 2 hours by Zoom | Optional" - "16 + 8 hours total per module | 32 + 16 hours total for the complete program" - "Module 1 is on Wednesdays 16:00-20:00, except the first session that will be on Monday at 16:00-20:00. Session dates: 7.9.26 (Monday), 16.9.26 (Wednesday), 23.9.26 (Wednesday), 7.10.26 (Wednesday)" - "Module 2 is on Mondays 16:00-20:00. Session dates: : 19.10.26 (Monday), 26.10.26 (Monday), 2.11 (Monday), 9.11.26 (Monday)" - "Practice sessions are on Thursdays at 18:00-20:00, via Zoom." - type: "image" alt: "Image 3" url: "https://www.graitechschool.co.il/wp-content/uploads/2024/06/%D7%A2%D7%9E%D7%A8%D7%95-scaled-e1718621125488.jpg" actions: [] - section: title: "Overview: Why this Program Exists" description: null elements: - type: "paragraph" text: "This program is built specifically for experienced product leaders who want to advance their skills as AI PM Builders, to safely and efficiently lead product development from concept to production." - type: "paragraph" text: "The role of the Product Manager has changed. PMs must leverage advanced AI capabilities and agentic engineering workflows to synthesize user research, generate product concepts, prototype interactive & testable products, define AI Metrics, perform Evals & define Guardrails, understand multi-step agents architecture and analyze agent Unit Economics, and lead strategic decision making before deployment." - type: "paragraph" text: "AI Products are probabilistic, are exceptionally difficult to test, expensive to scale, and heavily exposed to risks like runaway token loops, data exposure, and alignment failures. To ship successfully, PMs must track trajectory success metrics, engineer golden datasets for LLM-as-a-judge evaluations, hardcode agent circuit breakers, and break down complex token COGS into precise input, output, and retrieval buckets." - type: "image" alt: "Image 4" url: "https://www.graitechschool.co.il/wp-content/uploads/2022/12/Gilad-Kavalerchik__221220_146-e1671960198187.jpg" actions: [] - section: title: "AI PM Builder 1: Prototype & Validate" description: "This module provides hands-on skills on the strategic change in product teams, how to decide what to build and why to build it using automations, how to identify what to not automate, why AI-PRDs are essential, how design is changing, how to prototype with mock APIS, how to simulate backend logic, all while preventing friction with engineering." elements: - type: "text" label: "In-class sessions" text: "7.9.26 (Monday), 16.9.26, 23.9.26, 7.10.26 (Wednesdays 16:00-20:00)" - type: "text" label: "Practice Zoom sessions" text: "10.9.26, 17.9.26, 24.9.26, 1.10.26 (Thursdays 18:00-20:00)" - type: "text" label: "Core tools used" text: "Claude (claude.ai), Claude Code, Figma + Dev Mode, VS Code, Node.js, Mixpanel / Amplitude. Additional tools we will discuss but not dive into: Dovetail, Cursor, v0.dev, Bolt.new, Lovable." - type: "session" number: 1 title: "The AI PM Shift (7.9.26)" description: "Org Change, team friction, the “AI demo trap”, engineering handoff, data flywheels, product workflows, AI Value Mapping, Opportunity Discovery, Human-in-the-Loop Architecture." practice_zoom_session: "Practice Zoom session at 10.9.26" - type: "session" number: 2 title: "Automated Discovery & Design (16.9.26)" description: "Automated Qualitative Synthesis, AI Capability Scoping, AI-Native PRD Engineering, AI UX Design decisions (when to automate vs. require user confirmation, when to display model confidence metrics, how to expose system reasoning, how to implement user recovery controls)." practice_zoom_session: "Practice Zoom session at 17.9.26" - type: "session" number: 3 title: "Terminal Prototyping with Claude Code (23.9.26)" description: "Configuring Workspace, PRD-to-Code Mapping, AI Code-Generation Scripting, PM vs. Engineering Ownership, Prototype vs. production quality parameters, Avoiding organizational friction." practice_zoom_session: "Practice Zoom session at 24.9.26" - type: "session" number: 4 title: "Advanced Prototyping, Testing, and Handoff to developers (7.10.26)" description: "Embedding fake backend logic, Designing mock APIs, Configuring simulated real-time model behaviors, Handle failure states & system fallback routines, Code Refactoring & Auditing, Configuring Analytics, Handoff to developers." practice_zoom_session: "Practice Zoom session at 8.10.26" - type: "image" alt: "Image 5" url: "https://www.graitechschool.co.il/wp-content/uploads/2022/12/Gilad-Kavalerchik__221220_58-e1671953361203.jpg" actions: [] - section: title: "AI PM Builder 2: Evaluate & Ship" description: "This standalone module dives into AI Quality Metrics, Evals & Guardrails, Multi-step agents architecture, Agent Unit Economics, and strategic decision making before deployment. It highlights Human-in-the-loop automation limits and Agentic-by-default architecture in all sessions." elements: - type: "text" label: "In-class sessions" text: "19.10.26, 26.10.26, 2.11.26, 9.11.26 (Mondays 16:00-20:00)" - type: "text" label: "Practice Zoom sessions" text: "22.10.26, 29.10.26, 5.11.26, 12.11.26 (Thursdays 18:00-20:00)" - type: "text" label: "Core tools used" text: "Braintrust, Promptfoo, Lakera Guard, NeMo Guardrails, PostHog, Helicone, Anthropic Console / OpenRouter, TCO Calculators. Additional tools we will discuss but not dive into: Ragas, Guardrails AI, Galileo, Garak / PyRIT, Langfuse." - type: "session" number: 1 title: "AI Quality Metrics (19.10.26)" description: "Setting multi-layered performance metrics including leading and lagging indicators, cost-per-successful-task calculations and manual human-review rates, scoring definitions for precision, recall accuracy, and context faithfulness, Agentic Trajectory Mapping and Automated Release Gating." practice_zoom_session: "Practice Zoom session: 22.10.26" - type: "session" number: 2 title: "Evals & Golden Datasets (26.10.26)" description: "What does “good enough to ship” mean? Curating and formatting evaluation golden datasets, Configuring and tuning LLM-as-judge automations, Deploying automated evaluation runs into delivery pipelines, RAG Pipeline Evaluation, Agentic Trajectory Evals, and setting semantic tests to monitor compliance & tone." practice_zoom_session: "Practice Zoom session: 29.10.26" - type: "session" number: 3 title: "Guardrails, Agent Safety & Monitoring (2.11.26)" description: "How do you keep an AI product or agent safe and governed? Setting live input and output guardrail filtering, implementing agentic circuit breakers including token budget caps and systemic loop detection routines, setting agent permissions architecture, creating data fencing, set auditable compliance logging, and deploy real-time monitoring post-launch to detect semantic data drift or gradual model performance degradation." practice_zoom_session: "Practice Zoom session: 5.11.26" - type: "session" number: 4 title: "PM leadership on Model Selection, Agentic Unit Economics, and Production Readiness (9.11.26)" description: "How do PMs make the strategic calls on models, costs and margins? Model Optimization Architecture and the tradeoffs between commercial APIs vs. open-source, RAG vs. fine-tuning, SLMs vs. LLMs. Cost-Bucket Modeling, maximize margins using engineering levers, Agentic Financial Control including financial stress testing, Total Cost of Ownership (TCO) calculations, and Cross-Functional AI Governance." practice_zoom_session: "Practice Zoom session: 12.11.26" - type: "image" alt: "Image 6" url: "https://www.graitechschool.co.il/wp-content/uploads/2022/12/%D7%90%D7%93%D7%99%D7%A1%D7%9C%D7%90%D7%9D-%D7%95%D7%90%D7%9C%D7%97%D7%90%D7%9F-scaled.jpg" actions: [] - section: title: "What Participants Graduate With" description: null elements: - type: "paragraph" label: "Participants exit Module 1" text: "with the know-how to utilize Claude and Claude Code to accelerate discovery, generate machine-readable PRDs, construct functional prototypes, test the prototypes using multiple backend processes, and manage a successful handover to engineering." - type: "paragraph" label: "Participants exit Module 2" text: "with the know-how to determine whether an AI product is ready to be shipped, by establishing rigorous evaluation criteria, configuring safety guardrails, monitoring multi-step agent trajectories, and structuring optimized unit economic models to guarantee long-term profitability." actions: [] footer: links: [] info: [] social: []