smartsqlaimentor
Budget: $15 – $25 USD
Smart SQL AI Mentor is moving into development and I want the very first public release to focus exclusively on the SQL track. The goal is a full-stack, browser-based practice environment running on MySQL 8 where users can write, run and refine queries while an AI mentor guides them in real time.
Core feature set for this phase
• AI-generated schemas that refresh between sessions so every learner gets a unique, realistic dataset.
• Auto-drawn entity–relationship diagrams (ERDs) that update instantly when a new schema is spun up.
• Interactive query visualisation that shows execution flow, estimated costs and result previews side-by-side.
• AI query optimiser suggestions that flag inefficient joins, missing indexes or sub-optimal filters.
• AI Review summarising correctness, performance trade-offs and alternative approaches.
• “Smart” query visualisation layer that compares the user’s mental model with the actual execution path.
What I consider done
1. A working web interface wired to a containerised MySQL 8 instance.
2. All items above functioning end-to-end for at least three sample challenges.
3. Clean, well-documented code (frontend and backend) pushed to a Git repo with setup scripts.
4. Short Loom or text walkthrough explaining architecture choices and how to extend the challenge library.
Feel free to propose libraries or frameworks you believe will accelerate the build, but keep the stack mainstream so future contributors can ramp up quickly. I’ll review pull requests daily and can provide prompt feedback on UI/UX tweaks or model fine-tuning.
Core feature set for this phase
• AI-generated schemas that refresh between sessions so every learner gets a unique, realistic dataset.
• Auto-drawn entity–relationship diagrams (ERDs) that update instantly when a new schema is spun up.
• Interactive query visualisation that shows execution flow, estimated costs and result previews side-by-side.
• AI query optimiser suggestions that flag inefficient joins, missing indexes or sub-optimal filters.
• AI Review summarising correctness, performance trade-offs and alternative approaches.
• “Smart” query visualisation layer that compares the user’s mental model with the actual execution path.
What I consider done
1. A working web interface wired to a containerised MySQL 8 instance.
2. All items above functioning end-to-end for at least three sample challenges.
3. Clean, well-documented code (frontend and backend) pushed to a Git repo with setup scripts.
4. Short Loom or text walkthrough explaining architecture choices and how to extend the challenge library.
Feel free to propose libraries or frameworks you believe will accelerate the build, but keep the stack mainstream so future contributors can ramp up quickly. I’ll review pull requests daily and can provide prompt feedback on UI/UX tweaks or model fine-tuning.