Project Overview
MindlessML is a no-code machine learning platform that allows users to upload their datasets and train custom models without any technical knowledge. The goal is to make machine learning accessible to product teams, founders, and analysts who want to leverage their data without relying on specialized data science resources.
The platform abstracts the complexity of ML pipelines and provides a simple, guided experience from data ingestion to model deployment.
Key Features
- No-Code ML Workflow: Upload data, configure training, and deploy models without writing code
- Dataset Management: Secure upload and storage of user datasets with S3 integration
- Automated Model Training: Backend pipeline for training and evaluating ML models
- Simple UI Experience: Built with a focus on usability for non-technical users
- Scalable Architecture: Designed to support multiple concurrent training jobs and users
- Authentication & User Isolation: Each user manages their own secure workspace
Architecture Overview
MindlessML is built with a modern full-stack architecture designed for scalability and simplicity:
- Frontend: Next.js (React-based UI for user interaction and dashboard)
- Backend API: FastAPI (Python) for ML orchestration and business logic
- Database: Supabase (user management, metadata, and application state)
- Storage: AWS S3 (dataset and model artifact storage)
- ML Layer: Python-based training pipelines integrated into the backend
What I Built
- End-to-End SaaS Architecture: Designed and implemented the full system from frontend to ML pipeline orchestration
- Dataset Ingestion Flow: Built secure upload system integrated with S3 and backend validation
- ML Training Pipeline: Developed backend services to trigger, manage, and monitor model training jobs
- Authentication Layer: Implemented user-based access control and workspace isolation using Supabase
- Frontend Application: Built a clean and intuitive Next.js interface for managing datasets and models
- API Design: Structured RESTful API with FastAPI for scalable communication between frontend and ML services
Key Technologies Used
- Next.js
- TypeScript
- FastAPI (Python)
- Supabase
- AWS S3
- Python (ML pipelines)
- REST APIs
- Docker
Product Focus
MindlessML focuses on reducing the friction between data and machine learning by removing the need for engineering expertise. The platform is designed to help users quickly validate ideas, experiment with datasets, and deploy simple ML models in minutes.
Current Status
The platform is in its early stage, actively evolving with new features around automation, model explainability, and improved dataset handling.
Feedback and collaboration are always welcome.