5-Hour Agentic AI Course: Complete Interview Preparation
Mukul Raina
33,051 views
5-Hour Agentic AI Course: Complete Interview Preparation
===========
Timestamps:
===========
WHAT ARE AI AGENTS
0:00:00 Welcome & Course Overview
0:02:24 What Is an AI Agent (5 Properties)
0:04:36 The Agentic Spectrum
0:06:18 Agent Loop: Perceive-Reason-Act
0:09:18 Agent Components Overview
0:12:44 Case Study: Customer Support Agent
LLM FOUNDATIONS
0:15:11 LLM Foundations & Transformers
0:19:37 Context Window Mechanics
0:23:53 Temperature & Sampling Controls
0:26:51 LLM Limitations (5 Failure Modes)
0:30:11 Model Selection & Cost Trade-Offs
0:32:49 Section Summary & Reference Card
EXECUTION PATTERNS
0:35:07 Execution Patterns Intro
0:40:03 ReAct Pattern (Reason-Act-Observe)
0:43:10 Chain of Thought Reasoning
0:48:09 Prompt Chaining Pattern
0:49:14 Router Pattern (Intent Dispatch)
0:51:35 Reflection Pattern
0:54:39 Pattern Selection Mistakes
0:57:36 Section Summary & Reference Card
TOOL CALLING & FUNCTION USE
0:59:56 Tool Calling & Function Use Intro
1:04:40 Tool Schema Design
1:06:47 Tool Count Limits & Dynamic Loading
1:08:49 Structured Output & Validation
1:12:05 Tool Failure Handling & Retries
1:16:28 Parallel vs Sequential Tool Calls
1:19:21 MCP (Model Context Protocol)
1:26:21 Section Summary & Reference Card
CONTEXT, MEMORY & STATE
1:28:04 Context, Memory & State Intro
1:29:09 Short-Term Memory (Sliding Window)
1:34:05 Long-Term Memory (Persistent Store)
1:37:45 Context Window Budget Allocation
1:42:09 Reasoning Scratchpads
1:45:28 Common Memory Mistakes
RAG (RETRIEVAL-AUGMENTED GENERATION)
1:54:06 RAG Fundamentals
1:57:29 Chunking Strategies
1:59:28 Embeddings & Vector Search
2:06:34 Hybrid Search (BM25 + Semantic)
2:07:19 Re-Ranking for Precision
2:09:02 Few-Shot Examples for Behavior
2:12:21 RAG vs Fine-Tuning vs Long Context
2:16:14 Knowledge Base Sizing Walkthrough
2:20:47 Section Summary & Reference Card
MULTI-AGENT SYSTEMS
2:23:49 Multi-Agent Systems Intro
2:24:57 Supervisor-Worker Pattern
2:29:20 Agent Handoff & Escalation
2:34:30 Shared State & Coordination
2:42:53 When to Split vs Keep Single
RELIABILITY & GUARDRAILS
2:45:33 Reliability & Guardrails Intro
2:46:18 Loop Budgets & Termination
2:51:57 Graceful Degradation & Fallbacks
2:56:46 Idempotency Protection
SECURITY & PROMPT INJECTION
3:01:23 Security & Prompt Injection Intro
3:04:03 Input Filtering & Sanitization
3:05:54 Output Filtering & PII Detection
3:12:38 Authorization & Human-in-Loop
EVALUATION & TESTING
3:18:30 Evaluation & Testing Intro
3:20:22 Task Success & Quality Metrics
3:24:02 Regression Testing & Golden Sets
3:31:17 A/B Testing & Deployment
COST & LATENCY OPTIMIZATION
3:38:26 Cost & Latency Optimization Intro
3:43:15 Model Routing (Cost Tiering)
3:44:39 Semantic Caching
3:49:14 Latency Optimization & Streaming
3:53:11 Horizontal Scaling Patterns
3:59:41 Production Cost Projection
4:02:08 Section Summary & Reference Card
PRODUCTION ARCHITECTURE
4:05:42 Production Architecture Intro
4:07:30 Entry & Coordination Layers
4:09:38 Knowledge & State Layers
4:16:12 Full Interaction Trace Walkthrough
4:19:40 Failure Scenarios Walkthrough
4:22:09 Cost & Scaling Projections
4:24:42 25-Min Interview Presentation Guide
4:27:05 Common Interview Questions
4:32:15 Course Reference Card
HANDS-ON BUILD (PYTHON)
4:36:31 Hands-On Build: Agent from Scratch
4:41:10 Minimal ReAct Loop Implementation
4:45:53 Tool Definitions & Registry
4:49:28 Tool Execution & Error Handling
4:53:01 Memory Management Implementation
4:54:39 RAG Pipeline Implementation
4:56:19 Reliability Layer (Retries/Timeouts)
4:59:10 Idempotency & Loop Budgets
5:00:09 Authorization & Human Approval
5:04:01 Structured Logging & Tracing
5:06:28 Model Routing Implementation
5:07:57 Semantic Caching Implementation
5:08:58 Parallel Tool Execution
=========
About me:
=========
I'm Mukul Raina, a Senior Software Engineer and Tech Lead at Microsoft, with a Master's from the University of Oxford, UK. On this channel, I create technical deep dives on System Design, LeetCode, DSA, Data Engineering and ML/AI architectures.
#AgenticAI #AIAgents #LLM #RAG #SystemDesign #SoftwareEngineering #FAANG #TechInterview #InterviewPreparation #SeniorEngineer #ToolCalling #MultiAgentSystems #PromptEngineering #AIAgentsCourse
20 Comments
0:00:00 Welcome & Course Overview
0:02:24 What Is an AI Agent (5 Properties)
0:04:36 The Agentic Spectrum
0:06:18 Agent Loop: Perceive-Reason-Act
0:09:18 Agent Components Overview
0:12:44 Case Study: Customer Support Agent
LLM FOUNDATIONS
0:15:11 LLM Foundations & Transformers
0:19:37 Context Window Mechanics
0:23:53 Temperature & Sampling Controls
0:26:51 LLM Limitations (5 Failure Modes)
0:30:11 Model Selection & Cost Trade-Offs
0:32:49 Section Summary & Reference Card
EXECUTION PATTERNS
0:35:07 Execution Patterns Intro
0:40:03 ReAct Pattern (Reason-Act-Observe)
0:43:10 Chain of Thought Reasoning
0:48:09 Prompt Chaining Pattern
0:49:14 Router Pattern (Intent Dispatch)
0:51:35 Reflection Pattern
0:54:39 Pattern Selection Mistakes
0:57:36 Section Summary & Reference Card
TOOL CALLING & FUNCTION USE
0:59:56 Tool Calling & Function Use Intro
1:04:40 Tool Schema Design
1:06:47 Tool Count Limits & Dynamic Loading
1:08:49 Structured Output & Validation
1:12:05 Tool Failure Handling & Retries
1:16:28 Parallel vs Sequential Tool Calls
1:19:21 MCP (Model Context Protocol)
1:26:21 Section Summary & Reference Card
CONTEXT, MEMORY & STATE
1:28:04 Context, Memory & State Intro
1:29:09 Short-Term Memory (Sliding Window)
1:34:05 Long-Term Memory (Persistent Store)
1:37:45 Context Window Budget Allocation
1:42:09 Reasoning Scratchpads
1:45:28 Common Memory Mistakes
RAG (RETRIEVAL-AUGMENTED GENERATION)
1:54:06 RAG Fundamentals
1:57:29 Chunking Strategies
1:59:28 Embeddings & Vector Search
2:06:34 Hybrid Search (BM25 + Semantic)
2:07:19 Re-Ranking for Precision
2:09:02 Few-Shot Examples for Behavior
2:12:21 RAG vs Fine-Tuning vs Long Context
2:16:14 Knowledge Base Sizing Walkthrough
2:20:47 Section Summary & Reference Card
MULTI-AGENT SYSTEMS
2:23:49 Multi-Agent Systems Intro
2:24:57 Supervisor-Worker Pattern
2:29:20 Agent Handoff & Escalation
2:34:30 Shared State & Coordination
2:42:53 When to Split vs Keep Single
RELIABILITY & GUARDRAILS
2:45:33 Reliability & Guardrails Intro
2:46:18 Loop Budgets & Termination
2:51:57 Graceful Degradation & Fallbacks
2:56:46 Idempotency Protection
SECURITY & PROMPT INJECTION
3:01:23 Security & Prompt Injection Intro
3:04:03 Input Filtering & Sanitization
3:05:54 Output Filtering & PII Detection
3:12:38 Authorization & Human-in-Loop
EVALUATION & TESTING
3:18:30 Evaluation & Testing Intro
3:20:22 Task Success & Quality Metrics
3:24:02 Regression Testing & Golden Sets
3:31:17 A/B Testing & Deployment
COST & LATENCY OPTIMIZATION
3:38:26 Cost & Latency Optimization Intro
3:43:15 Model Routing (Cost Tiering)
3:44:39 Semantic Caching
3:49:14 Latency Optimization & Streaming
3:53:11 Horizontal Scaling Patterns
3:59:41 Production Cost Projection
4:02:08 Section Summary & Reference Card
PRODUCTION ARCHITECTURE
4:05:42 Production Architecture Intro
4:07:30 Entry & Coordination Layers
4:09:38 Knowledge & State Layers
4:16:12 Full Interaction Trace Walkthrough
4:19:40 Failure Scenarios Walkthrough
4:22:09 Cost & Scaling Projections
4:24:42 25-Min Interview Presentation Guide
4:27:05 Common Interview Questions
4:32:15 Course Reference Card
HANDS-ON BUILD (PYTHON)
4:36:31 Hands-On Build: Agent from Scratch
4:41:10 Minimal ReAct Loop Implementation
4:45:53 Tool Definitions & Registry
4:49:28 Tool Execution & Error Handling
4:53:01 Memory Management Implementation
4:54:39 RAG Pipeline Implementation
4:56:19 Reliability Layer (Retries/Timeouts)
4:59:10 Idempotency & Loop Budgets
5:00:09 Authorization & Human Approval
5:04:01 Structured Logging & Tracing
5:06:28 Model Routing Implementation
5:07:57 Semantic Caching Implementation
5:08:58 Parallel Tool Execution
Can you also please share the notion link as well, if possible?
5-Hour Agentic AI Course: Complete Interview Preparation
Mukul Raina
33,051 views
5-Hour Agentic AI Course: Complete Interview Preparation
===========
Timestamps:
===========
WHAT ARE AI AGENTS
0:00:00 Welcome & Course Overview
0:02:24 What Is an AI Agent (5 Properties)
0:04:36 The Agentic Spectrum
0:06:18 Agent Loop: Perceive-Reason-Act
0:09:18 Agent Components Overview
0:12:44 Case Study: Customer Support Agent
LLM FOUNDATIONS
0:15:11 LLM Foundations & Transformers
0:19:37 Context Window Mechanics
0:23:53 Temperature & Sampling Controls
0:26:51 LLM Limitations (5 Failure Modes)
0:30:11 Model Selection & Cost Trade-Offs
0:32:49 Section Summary & Reference Card
EXECUTION PATTERNS
0:35:07 Execution Patterns Intro
0:40:03 ReAct Pattern (Reason-Act-Observe)
0:43:10 Chain of Thought Reasoning
0:48:09 Prompt Chaining Pattern
0:49:14 Router Pattern (Intent Dispatch)
0:51:35 Reflection Pattern
0:54:39 Pattern Selection Mistakes
0:57:36 Section Summary & Reference Card
TOOL CALLING & FUNCTION USE
0:59:56 Tool Calling & Function Use Intro
1:04:40 Tool Schema Design
1:06:47 Tool Count Limits & Dynamic Loading
1:08:49 Structured Output & Validation
1:12:05 Tool Failure Handling & Retries
1:16:28 Parallel vs Sequential Tool Calls
1:19:21 MCP (Model Context Protocol)
1:26:21 Section Summary & Reference Card
CONTEXT, MEMORY & STATE
1:28:04 Context, Memory & State Intro
1:29:09 Short-Term Memory (Sliding Window)
1:34:05 Long-Term Memory (Persistent Store)
1:37:45 Context Window Budget Allocation
1:42:09 Reasoning Scratchpads
1:45:28 Common Memory Mistakes
RAG (RETRIEVAL-AUGMENTED GENERATION)
1:54:06 RAG Fundamentals
1:57:29 Chunking Strategies
1:59:28 Embeddings & Vector Search
2:06:34 Hybrid Search (BM25 + Semantic)
2:07:19 Re-Ranking for Precision
2:09:02 Few-Shot Examples for Behavior
2:12:21 RAG vs Fine-Tuning vs Long Context
2:16:14 Knowledge Base Sizing Walkthrough
2:20:47 Section Summary & Reference Card
MULTI-AGENT SYSTEMS
2:23:49 Multi-Agent Systems Intro
2:24:57 Supervisor-Worker Pattern
2:29:20 Agent Handoff & Escalation
2:34:30 Shared State & Coordination
2:42:53 When to Split vs Keep Single
RELIABILITY & GUARDRAILS
2:45:33 Reliability & Guardrails Intro
2:46:18 Loop Budgets & Termination
2:51:57 Graceful Degradation & Fallbacks
2:56:46 Idempotency Protection
SECURITY & PROMPT INJECTION
3:01:23 Security & Prompt Injection Intro
3:04:03 Input Filtering & Sanitization
3:05:54 Output Filtering & PII Detection
3:12:38 Authorization & Human-in-Loop
EVALUATION & TESTING
3:18:30 Evaluation & Testing Intro
3:20:22 Task Success & Quality Metrics
3:24:02 Regression Testing & Golden Sets
3:31:17 A/B Testing & Deployment
COST & LATENCY OPTIMIZATION
3:38:26 Cost & Latency Optimization Intro
3:43:15 Model Routing (Cost Tiering)
3:44:39 Semantic Caching
3:49:14 Latency Optimization & Streaming
3:53:11 Horizontal Scaling Patterns
3:59:41 Production Cost Projection
4:02:08 Section Summary & Reference Card
PRODUCTION ARCHITECTURE
4:05:42 Production Architecture Intro
4:07:30 Entry & Coordination Layers
4:09:38 Knowledge & State Layers
4:16:12 Full Interaction Trace Walkthrough
4:19:40 Failure Scenarios Walkthrough
4:22:09 Cost & Scaling Projections
4:24:42 25-Min Interview Presentation Guide
4:27:05 Common Interview Questions
4:32:15 Course Reference Card
HANDS-ON BUILD (PYTHON)
4:36:31 Hands-On Build: Agent from Scratch
4:41:10 Minimal ReAct Loop Implementation
4:45:53 Tool Definitions & Registry
4:49:28 Tool Execution & Error Handling
4:53:01 Memory Management Implementation
4:54:39 RAG Pipeline Implementation
4:56:19 Reliability Layer (Retries/Timeouts)
4:59:10 Idempotency & Loop Budgets
5:00:09 Authorization & Human Approval
5:04:01 Structured Logging & Tracing
5:06:28 Model Routing Implementation
5:07:57 Semantic Caching Implementation
5:08:58 Parallel Tool Execution
=========
About me:
=========
I'm Mukul Raina, a Senior Software Engineer and Tech Lead at Microsoft, with a Master's from the University of Oxford, UK. On this channel, I create technical deep dives on System Design, LeetCode, DSA, Data Engineering and ML/AI architectures.
#AgenticAI #AIAgents #LLM #RAG #SystemDesign #SoftwareEngineering #FAANG #TechInterview #InterviewPreparation #SeniorEngineer #ToolCalling #MultiAgentSystems #PromptEngineering #AIAgentsCourse