Skip to content

Commit c9d2c75

Browse files
add draft Agentic AI workshop content to repo
1 parent 5f88202 commit c9d2c75

File tree

6 files changed

+237
-0
lines changed

6 files changed

+237
-0
lines changed

workshop/README.md

Lines changed: 71 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,71 @@
1+
# AI-Driven Development Workflows
2+
3+
## Speed Up Development with Intelligent Automation
4+
5+
Welcome to the AI-Driven Development Workflows module! Now that you have deployed Coder and created intelligent templates, it's time to experience the future of software development. This module demonstrates how AI transforms every aspect of the development lifecycle—from initial code generation to deployment and monitoring.
6+
7+
### What We'll Accomplish
8+
9+
In this module, you will:
10+
11+
1. **Learn AI-Powered Coding Workflows** - Experience intelligent code generation, completion, and refactoring with Amazon Q Developer and AWS Bedrock
12+
2. **Create Intelligent Deployment Automation** - Deploy applications using AI-powered infrastructure provisioning and optimization
13+
3. **Build Agentic Development Workflows** - Create autonomous development agents that handle routine tasks and complex problem-solving
14+
15+
## The AI-First Development Philosophy
16+
17+
AI-driven development isn't just about code completion—it's a fundamental shift in how we approach software engineering:
18+
19+
### Traditional Development vs. AI-Driven Development
20+
21+
| Traditional Approach | AI-Driven Approach |
22+
|---------------------|--------------------|
23+
| Manual code writing | AI-assisted generation with human guidance |
24+
| Reactive debugging | Proactive issue detection and resolution |
25+
| Manual testing | Intelligent test generation and execution |
26+
| Static documentation | Dynamic, context-aware documentation |
27+
| Manual code reviews | AI-enhanced reviews with security scanning |
28+
| Reactive monitoring | Predictive analytics and automated remediation |
29+
30+
### Key Principles
31+
32+
- **Human-AI Collaboration**: AI amplifies human creativity rather than replacing it
33+
- **Context-Aware Assistance**: AI understands your codebase, patterns, and objectives
34+
- **Continuous Learning**: Systems improve based on team patterns and feedback
35+
- **Proactive Intelligence**: AI anticipates needs and suggests optimizations
36+
- **End-to-End Automation**: AI workflows span the entire development lifecycle
37+
38+
## Architecture Overview
39+
40+
Our AI-driven development platform integrates multiple AWS AI services with Coder:
41+
42+
![Coder AI Architecture diagram](./images/AWSCoderAgenticAI.png)
43+
44+
## Real-World AI Development Scenarios
45+
46+
### Scenario 1: New Feature Development
47+
**Traditional Time**: Weeks
48+
**AI-Driven Time**: Days
49+
50+
1. **AI Requirements Analysis**: Natural language feature description → detailed technical requirements
51+
2. **Intelligent Code Generation**: AI generates boilerplate, API endpoints, and database schemas
52+
3. **Automated Testing**: AI creates comprehensive test suites based on requirements
53+
4. **Smart Code Review**: AI identifies potential issues, security vulnerabilities, and optimization opportunities
54+
5. **Intelligent Deployment**: AI optimizes infrastructure configuration and deployment strategy
55+
56+
### Module Learning Objectives
57+
58+
By the end of this module, you will be able to:
59+
60+
**AI-Powered Development Skills:**
61+
- Leverage Amazon Q Developer for intelligent code completion and generation
62+
- Use AWS Bedrock/Anthropic Claude Code for advanced code review and architectural guidance
63+
- Implement AI-driven refactoring and optimization workflows
64+
- Create context-aware development assistants
65+
66+
> **💡 AI Development Mindset**: Think of AI as your intelligent pair programming partner. The goal is human-AI collaboration that amplifies creativity and productivity.
67+
68+
> **⚠️ Warning**: The examples and sample code provided in this workshop are intended to be consumed as instructional content. These examples are not intended for use in production environments.
69+
70+
### [Ready to Experience the Future?](./modules/ai-driven-development.md)
71+
Let's dive into [AI-powered development](./modules/ai-driven-development.md) workflows and transform how you build software. The future of development is intelligent, automated, and incredibly productive!
125 KB
Loading
382 KB
Loading
108 KB
Loading
Lines changed: 42 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,42 @@
1+
# AI-Driven Task Automation Workflows
2+
3+
## Transforming Development Tasks with Intelligent Automation
4+
5+
AI-driven automation workflows represent a potentially exponential boost to developer productivity by enabling Agentic AI to autonomously perform routine development tasks with minimal input and supervision. Instead of using precious Developer cycles, AI can automate standard SDLC tasks, allowing developers to focus on creative problem-solving and innovation.
6+
7+
### Workflow 2: AI-Driven Task Automation with Claude Code
8+
#### Scenario: Perform an automated Well-Architected Review, and act on any findings
9+
10+
Let's create a Coder Task that will perform a Well-Architected review of the Web App we previously created.
11+
12+
#### Step 1: Create Your Task-Automation Workspace
13+
14+
Create a Task using the AWS Workshop - Kubernetes with Claude Code template:
15+
1. **Access your Coder dashboard** and click "Tasks"
16+
2. **Within the Task UI, Select the AWS Workshop - Kubernetes with Claude Code template from the drop-down** (created in the previous module)
17+
3. **Configure the Task prompt**:
18+
```bash
19+
Analyze the Task Management Web App found at https://github.com/your-git-id/ai-dev-workflows.git, perform a Well Architected Review of the application with a focus on the Security pillar. Create up to two additional Coder workspaces using the AWS Workshop - Kubernetes with Claude Code task template. Use the issues identified as the Task prompt for each additional workspace, and ensure the Task prompt for the new workspaces specifies that updates should be made as PRs to the original git repo at https://github.com/your-git-id/ai-dev-workflows.git
20+
```
21+
22+
4. **Click "Run task"** and wait for it to start
23+
24+
> **ℹ️ Info**: The Coder Task UI will automatically provision a task-based workspace and Claude Code will begin analyzing the provided Task prompt.
25+
26+
The Claude Code Web UI in the left pane will begin by creating and updating a "To Do List" of activities to be performed. As it progresses, you can monitor the Agent's actions. It will most likely prompt for your approval and direction on how to move ahead to create new workspaces to resolve the identified issues. Depending on how you respond, the Agent will spawn up to two additional Tasks to remediate the findings.
27+
28+
![Coder AI Driven Review](../images/ai-driven-app-review.png)
29+
30+
#### Step 2: Monitor and review Async Task-Automation Workspaces
31+
32+
Once your initial Task completes:
33+
34+
1. **Re-open the Tasks** from your Coder dashboard
35+
2. **Review the new Tasks** created by your initial Task-Automation workflow
36+
3. **Open the new tasks** and monitor the changes made, and optionally submit PR's to your git repo
37+
38+
![Coder AI Driven Tasks](../images/ai-driven-task-automation.png)
39+
40+
Take some time on your own and explore each of the created Tasks and evaluate the results of the remediation activities performed by Claude Code. Experiment by instructing Claude Code to unit test the changes and/or smoke-test a deployment to your AWS account.
41+
42+
> **🚀 Workflow Automation**: These AI automation workflows can reduce time spent on routine activities, while dramatically boosting developer productivity. Start by automating one workflow and experimenting until you get consistent results. Gradually add more AI automation workflows as your team becomes comfortable with specific Agent capabilities and required prompts to get the desired results.
Lines changed: 124 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,124 @@
1+
# AI-Driven Development
2+
3+
## Speed Up Development with Intelligent Automation
4+
5+
AI-driven development workflows represent a fundamental shift from reactive to proactive development. Instead of waiting for issues to arise, AI anticipates needs, suggests optimizations, and automates routine tasks, allowing developers to focus on creative problem-solving and innovation.
6+
7+
### Setting Up Your AI Development Environment
8+
9+
Let's create your Coder workspace with comprehensive AI development tools.
10+
11+
#### Step 1: Access Your AI-Enhanced Workspace
12+
13+
Create a workspace using the AWS Workshop - EC2 (Linux) Q Developer template:
14+
1. **Access your Coder dashboard** and click "Create Workspace"
15+
2. **Select the AWS Workshop - EC2 (Linux) Q Developer template** (created from the [template](../../templates/awshp-linux-q-base/README.md) in this repo)
16+
3. **Configure the workspace parameters**:
17+
- **Name**: `linux-qdev-workspace`
18+
- **Instance type**: 2 vCPU, 4 GiB RAM
19+
- **Region**: us-east-1 (Default)
20+
- **Disk Size**: 30 GB (Default)
21+
22+
4. **Click "Create Workspace"** and wait for it to start
23+
24+
> **ℹ️ Info**: The selected Coder workspace template will automatically provision the AWS CLI, CDK, Amazon Q Developer CLI and other tools needed for AI-Driven AWS Development.
25+
26+
#### Step 2: Access Your Linux Q Developer Workspace
27+
28+
Once your workspace is running:
29+
30+
1. **Open the workspace** from your Coder dashboard
31+
2. **Launch code-server** or your preferred editor
32+
3. **Open a terminal** within the workspace
33+
34+
#### Step 3: Initialize Workshop Git/Github repo
35+
Once in your workspace, let's create a workshop directory and initialize a git repo:
36+
```bash
37+
mkdir ai-dev-workflows
38+
```
39+
Now from code-server or VS Code:
40+
41+
1. **Use File/Open Folder** to open the workshop directory
42+
2. **Use Git extension** to initialize a git repository in the current directory
43+
44+
#### Step 4: Initialize AI Development Tools
45+
Back in your workspace terminal session, let's set up the AI development environment:
46+
```bash
47+
# Initialize the Q Developer CLI
48+
q login # Use for Free with Builder ID option, and follow prompts
49+
q chat # Initialize chat session
50+
```
51+
52+
### Workflow 1: AI-Assisted Feature Development
53+
#### Scenario: Create a simple Cloud-Native Task Management Web App
54+
Let's walk through developing a new feature using AI assistance from start to finish.
55+
56+
Step 1: Requirements Analysis with AI, start by describing your feature in natural language:
57+
```bash
58+
# Use Amazon Q Developer to analyze requirements with the following prompt:
59+
analyze the following requirements: "Create a simple task management web app that tracks task id, description, priority, and completion date. Provide two ways to interact with the data, one that summarizes open tasks by priority and another lists completed tasks by date"
60+
```
61+
Amazon Q will provide:
62+
- Technical requirements breakdown
63+
- Architecture suggestions
64+
- Implementation approach
65+
- Potential challenges and solutions
66+
67+
Step 2: AI-Generated Project Structure
68+
```bash
69+
# Next, have Amazon Q generate the supporting project structure with the following prompt:
70+
generate a supporting project structure for an AWS CDK application that uses typscript for the front end components and python for back-end API components
71+
```
72+
73+
> **ℹ️ Info**: Notice how Amazon Q always prompts you to "trust" it when creating or updating content in your Coder workspace.
74+
75+
This should create something similar to this:
76+
```bash
77+
task-management-app/
78+
├── infrastructure/ # AWS CDK TypeScript code
79+
│ ├── bin/app.ts # CDK app entry point
80+
│ ├── lib/ # CDK stack definitions
81+
│ │ ├── database-stack.ts # DynamoDB table
82+
│ │ ├── backend-stack.ts # Lambda + API Gateway
83+
│ │ └── frontend-stack.ts # S3 + CloudFront
84+
│ └── package.json # CDK dependencies
85+
├── backend/ # Python Lambda functions
86+
│ ├── src/
87+
│ │ ├── models/ # Data models
88+
│ │ ├── services/ # Business logic
89+
│ │ └── handlers/ # Lambda handlers
90+
│ └── requirements.txt # Python dependencies
91+
├── frontend/ # React TypeScript app
92+
│ ├── src/
93+
│ │ ├── components/ # React components
94+
│ │ ├── services/ # API client
95+
│ │ └── types/ # TypeScript interfaces
96+
│ └── package.json # React dependencies
97+
└── scripts/ # Deployment scripts
98+
```
99+
100+
Step 3: AI-Generated AWS Deployment
101+
```bash
102+
# Smoke-test deployment to AWS by having Amazon Q deploy the generated web app to the current AWS account with the following prompt:
103+
Smoke test the web app deployment to the current AWS account using the created deployment scripts
104+
```
105+
106+
> **ℹ️ Info**: Notice how Amazon Q will find and debug issues as it works with the existing scripts and workspace environment, installing required dependencies as needed. Additionally, you will most likely see Amazon Q iterate across Lambda Functions, Back-End Schema, and other component issues as it tests the CDK stacks being deployed.
107+
108+
When completed, at least your Database and Backend stacks should be successfully deployed and smoke-tested. You can continue to prompt Amazon Q to complete the full application deployment, if desired. It is suggested you commit and push changes to your workshop Git repo at this point, as this Git repo will be used in the next AI-Driven Workflow example.
109+
110+
Step 4: Cleanup AI-Generated AWS Deployment
111+
```bash
112+
# Have Amazon Q safely remove any deployments created for smoke-testing from the current AWS account with the following prompt:
113+
Remove any CDK stack deployments used for smoke-testing the task mananagement web app from the current AWS account. Double-check that only task management stacks are being deleted and nothing else.
114+
```
115+
This should remove any deployed components and ensure Amazon Q double-checks and reviews what was deleted. You can now end your Q CLI chat session with:
116+
```bash
117+
/quit
118+
```
119+
120+
> **🚀 Workflow Optimization**: These AI development workflows can reduce development time by 60-80% while improving code quality. Start with one workflow and gradually add more as your team becomes comfortable.
121+
122+
## [Next Steps](ai-driven-automation.md)
123+
124+
Now that you've experimented with AI-Driven Development, you can now see how [AI-Driven Automation](ai-driven-automation.md) can support your development workflow.

0 commit comments

Comments
 (0)