Follow the fourteen-part route on the course hub. This index also retains the earlier reference library; a longer list is not a requirement to read every page before doing useful work.
Getting started
Getting Started with AI for VLSI Engineers
Useful AI Tasks: Start Small and Checkable
Assistants, Models, Tools, and Your Responsibility
Your First RTL Explanation
Improve the Answer and Check It Independently
Choose a Tool for the Task
Just enough AI
Just Enough AI: Models and Context
What a Language Model Does
Next-Token Prediction and the Sampling Loop
Training and Inference
Parameters at an Intuitive Level
Fluency, Correctness and Knowledge Cutoff
Temperature and Sampling
Tokens and Tokenization
Tokens, Words and Characters
Subword Tokens and Byte-Pair Encoding
Why Some Text Costs More Tokens
Approximate Rules of Thumb
Estimating Tokens Before Sending
Context Window and Attention for Users
What Counts Against the Window
When the Window Fills
Attention in One Picture
Long Inputs and the Lost Middle
A Context-Budget Planner
Asking clearly
Prompting
Anatomy of a Prompt
Five Parts of a Prompt
Roles: System, User, and Assistant
Specific and Parseable
Weak Versus Strong Prompts
Prompt Template Builder
Examples and Reasoning
Zero-Shot and Few-Shot Prompts
Choosing and Ordering Examples
Step-by-Step Reasoning Prompts
Self-Checks and Structured Output
Prompt Debugging
Diagnose the Failure First
Change One Thing at a Time
Keep a Small Set of Test Prompts
A Prompt Regression Harness
Verification and safe use
Trust and Verification
Hallucination and Failure Modes
Invented APIs, Flags and Signal Names
Plausible but Wrong Numbers
Overconfident Tone and Sycophancy
Instruction Drift in Long Sessions
Subtle Bugs That Compile
Verifying AI Output
The Verify-First Checklist
Differential Testing Against a Reference
Property-Based Checks
Review the Diff, Not Just the Result
When to Refuse to Trust
Evaluating AI Tools for Your Team
Define the Task Before Measuring Anything
Build a Small Labelled Test Set
Measure Pass Rate and Time Saved
Track Cost per Accepted Result and Decide
Everyday engineering assistance
Using AI for Engineering Work
RTL and Testbench Assistance
Explaining Unfamiliar RTL
Drafting RTL from a Spec
Drafting Testbenches
Reviewing for Latches, Widths and Assertions
Assumptions and Acceptance
Debugging with AI
Log Slices That Help
Ranked Hypotheses with Evidence
Summarising and Clustering Logs
The Engineer Runs the Experiment
Scripts and Automation
Reviewing Generated Scripts
Idempotence and Dry-Run Flags
Testing a Script on a Copy
A Small Static Checker for Scripts
Data, Privacy and IP
What Never to Paste
Policy and Approved Tools
Sanitise Before Sharing
Local Tools, Hosted Tools, and Security Review
Role-based VLSI walkthroughs
Engineering Walkthroughs
Specification to Requirements
RTL Review
Verification Plan and Assertions
Scoreboard Debug
Timing Report
Physical Design Reports
Engineering Writing
Scripts, APIs and working copies
Practical Foundations for AI Workflows
APIs: Requests, Responses and Errors
JSON Records and Meaningful Validation
Repositories, Diffs and Isolated Changes
Working with Waveforms and Report Images
Engineering data
Engineering Data
Files, CSV, and JSON
Tables and IDs
Basic SQL for Engineering Questions
Regression Database Lab
Queries Versus Search
RAG and grounded answers
Retrieval-Augmented Generation
The Retrieval Workflow
Preparing Engineering Documents
Keyword and Semantic Search
Embeddings and Vector Storage
Grounded Answers and Abstention
Specification Assistant Lab
Tools, workflows and agents
Agents and Tool Use
What an Agent Is
Prompts, Scripts, and Agents
The Agent Loop Step by Step
Agent Components
Agent Loop Simulator
When a Script Wins
Tool Use and Structured Output
Describing Tools to a Model
Propose, Execute, Validate
Handling Malformed Calls
Idempotent and Side-Effecting Tools
Tool-Call Validator Lab
Planning, Memory and Stopping
Task Decomposition and Success Criteria
Memory Outside the Model and Summaries
Step and Cost Budgets
Loop Detection and Stopping
Multi-Agent Patterns and When Not to Use Agents
Planner, Worker and Reviewer Patterns
Hand-off Cost and Error Compounding
Complexity and Debuggability
Script, Workflow or Agent
Design a simple agent
Agent Design
Define the Task and Boundaries
Design the Architecture
Build Safe Read-Only Tools
Control the Decision Loop
Record State and Hand Over
Handle Errors and Budgets
Build the Starter Controller
Evaluate Your Assistant
MCP and reusable skills
MCP and Skills
What MCP Is
Clients, Servers, and Capabilities
Connect Read-Only Tools
Permissions and Untrusted Content
Reusable Engineering Skills
One-Tool Server Extension
Small multi-agent workflows
Multi-Agent Design
When to Split a Task
Roles and Task Contracts
Context and Handoffs
Merge and Disagreements
Permissions, Budgets and Failures
Review Team Lab
Troubleshooting and final project
Practical Capstone
Define the Assistant
Assemble the Workflow
Troubleshoot the Workflow
Evaluate Known and Unseen Cases
Compare Effort and Cost
Final Review and Handover
Optional reference library
Tokens, Cost and Latency
Cost Arithmetic
The Cost Formula
A Code-Review Job
A Nightly Regression Summary
Sensitivity and a Cost Calculator
Latency and Throughput
Time to First Token and Decode Speed
A Simple Latency Model
Streaming, Batching and Interactive Use
Rate Limits and Retries
Reducing Tokens
Prefix Caching
Summarisation and Information Loss
Retrieval Instead of Pasting
Trimming Logs and RTL
Agentic Workflows for Chip Teams
Regression Triage Agent
The Nightly Triage Loop
Clustering Failures by Signature
Ranking Suspects with Evidence
Read-Only by Default
Evaluating Against Labelled History
Lint-and-Fix Loops
Anatomy of the Loop
Acceptance Rules and the Test Suite
Limits and Failure Modes
A Loop Simulator with a Fake Fixer
Guardrails and Human in the Loop
Permission Tiers and Least Privilege
Sandboxing and Approval Gates
Rate, Cost Limits and Audit Logs
Secret Handling
Prompt Injection from Files and Logs
Evaluating and Operating Agents
Task Success Rate on a Held-Out Set
Cost and Latency per Task
Regression Suites for Agent Changes
Drift, Incident Review, and Shutoff Rules
Interview and Judgment
AI Interview Questions
Tokens and Context Questions
Cost Estimation Questions
Prompting Questions
Verification Questions
Agent Questions
Guardrails and Judgment Questions
Design Scenarios with Numbers
Estimate a Review Assistant Budget
Design a Regression-Triage Agent
Give a Model a 300k-Token Codebase
Evaluate Before Adopting a Tool
Related topics