Agentic AI Test Case Generator

I am looking to build an **LLM-powered Agentic AI system to autonomously generate and execute test cases for a serverless application**.

My backend runs on **AWS Lambda**, with all infrastructure defined using **AWS CDK (TypeScript)**. The goal is to introduce **AI-driven automation into the quality-assurance pipeline**, where an intelligent agent can analyze application workflows and dynamically generate executable test cases.

**Objective**
The agent should be capable of understanding application logic, identifying key workflows such as authentication and data input validation, and automatically generating comprehensive test scenarios that can be executed within our CI pipeline.

**Scope**

* The agent should discover and model key workflows, including:

* Sign-in / sign-up authentication paths
* Data-entry validation checkpoints
* It should automatically generate **edge-case inputs** for:

* Text fields
* File uploads
* API request and response handling
* Generated tests should integrate directly with the CI pipeline and run as:

* **Python (pytest)** or
* **JavaScript (Jest)** test suites
* The orchestration layer can use **LangChain, LangGraph, or a comparable framework**, provided it integrates well with AWS.
* The solution should integrate with existing **AWS Lambda functions** and ideally be able to provision temporary testing environments using **AWS CDK**, ensuring tests never run against production resources.

**Architecture & Design Expectations**

* Provide a **clear software architecture design** for the AI agent and testing pipeline.
* Define how the agent analyzes workflows, generates test scenarios, and integrates with CI/CD.
* Explain how AWS components (Lambda, CDK, storage, and orchestration services) will interact with the agent.
* Ensure the architecture is **scalable, secure, and maintainable** for future workflows.

**Deliverables**

1. Source code for the AI agent with clear structure and documentation.
2. Architecture design and documentation explaining system components and workflow.
3. AWS CDK constructs or deployment documentation to run the agent in my AWS environment.
4. Sample generated test suites covering:

* Authentication success and failure scenarios
* Text input edge cases
* File upload validations
* API request and response edge cases
5. A README or short walkthrough (Loom/video or documentation) explaining how the system works and how it can be extended as new workflows are added.

If you have prior experience building **autonomous testing systems, AI agents, or workflow-aware automation tools**, I would be happy to discuss the approach and milestones. I am available to provide quick feedback during development to help iterate efficiently.