I Built an AI-Powered Daily LeetCode Planner Using Workflow Automation

Recently, I explored workflow automation using Kestra and ended up building a small but powerful system, an AI-powered Daily LeetCode Planner.

The goal was simple:

Stop wasting time deciding what to solve every day and automate the entire process.

What it does

Every morning, the system automatically:

• Fetches coding problems from LeetCode
• Selects a balanced set of Easy, Medium, and Hard questions
• Sends them to an AI model for structured analysis
• Generates a full study plan with patterns and hints
• Emails it directly to me

Tech Stack

• Kestra (workflow orchestration)
• Python (data processing and logic)
• Google Gemini API (AI analysis)
• LeetCode GraphQL API (problem data source)
• SMTP (email automation)

How the workflow works

1. Fetch problems

Pulls a list of 50 problems from LeetCode GraphQL API.

2. Select questions

A Python script:

• Filters valid entries
• Groups by difficulty
• Randomly selects a balanced set

3. AI analysis

Gemini processes the selected questions and generates:

• Solving patterns
• Important concepts
• One-line hints
• Interview insights
• Recommended solving order

4. Email output

The final structured plan is sent automatically to my email every morning.

Why I built this

As someone practicing DSA regularly, I noticed:

• Time wasted choosing problems
• No structured daily flow
• No pattern-based learning system

So I automated it instead.

What I learned

This project taught me:

• Workflow automation can connect full systems, not just scripts
• AI is best used for structuring and reasoning, not just answers
• Small automation pipelines can quickly become powerful tools
• Debugging multi-step workflows is a different skill than normal coding

Flow YAML File (Github): https://github.com/rishabhrawat05/AI-Daily-LeetCode-Planner-Kestra