An end-to-end ML workflow automation tool powered by Groq's Llama models — from research to deployment, all in one place.
ML Engineer is an AI-powered assistant that handles the full machine learning pipeline for you. Instead of stitching together multiple tools and scripts, you get a single interface that:
- Searches and ranks research papers from arXiv
- Analyzes your dataset automatically
- Recommends the best ML algorithm for your data (with reasoning)
- Trains models in one click with real-time progress updates
- Generates professional reports with charts and insights
Think of it as your personal ML co-pilot — you bring the data and the ideas, it handles the heavy lifting.
Note: All AI features require your own Groq API key. No fallbacks or mock responses are used.
┌─────────────────────────────────────────────────────────┐
│ Frontend (React) │
│ - Futuristic UI with animations │
│ - Paper Search, Data Lab, Model Engine, Insights │
└──────────────────────┬──────────────────────────────────┘
│ HTTP & WebSockets
┌──────────────────────▼──────────────────────────────────┐
│ Backend (FastAPI) │
│ - File Uploads, WebSocket Streaming, Experiment Storage│
└──────────────────────┬──────────────────────────────────┘
│
┌──────────────┴──────────────┐
│ │
┌───────▼───────┐ ┌──────▼───────┐
│ LLM Engine │ │ ML Pipeline │
│ - Groq Llama │ │ - Scikit-learn│
│ - Local Ollama│ │ - Preprocessing│
└───────────────┘ └──────────────┘
The frontend (React) communicates with the backend (FastAPI) over HTTP and WebSockets. The backend connects to two subsystems: the LLM engine (Groq or a local Ollama instance) and the ML pipeline (scikit-learn for training and preprocessing).
| Feature | What It Does |
|---|---|
| Paper Search | Search arXiv and get AI-ranked results relevant to your topic |
| Data Lab | Upload a CSV and get instant statistics, missing value detection, and data profiling |
| AI Recommendation | Receive a suggested algorithm for your dataset, with an explanation of why |
| Auto-Train | Train a model with one click; watch real-time progress via WebSocket |
| Insights | Download a full ML report with charts and next-step recommendations |
Follow these steps in order. It should take less than five minutes.
If you don't have one yet, sign up at console.groq.com and create a free API key.
In the root directory of this project, create a file named .env and add the following line:
GROQ_API_KEY=gsk_your_actual_key_hereReplace gsk_your_actual_key_here with your real key. This file is never committed to version control.
Open a terminal, navigate to the backend folder, and run:
cd backend
python -u main.pyThe backend will start on http://localhost:8000.
Open a second terminal, navigate to the frontend folder, and run:
cd frontend
npm install # only needed the first time
npm run devThe frontend will start on http://localhost:5173.
Visit http://localhost:5173 in your browser. You're ready to go.
A sample dataset is included at data/sample.csv. Use it to walk through the full workflow end-to-end without needing your own data:
- Upload
sample.csvin the Data Lab tab - Review the auto-generated data profile
- Click Get AI Recommendation to see which algorithm fits best
- Hit Train and watch the model build in real time
- Head to Insights to download your report
| Layer | Technology |
|---|---|
| Frontend | React, Vite, Framer Motion |
| Backend | FastAPI, SQLite |
| AI / LLM | Groq Llama 3.3-70B, Local Ollama (optional) |
| ML | Scikit-learn, Pandas, NumPy |
The backend won't start.
Make sure Python 3.9+ is installed and your .env file is in the root directory (not inside backend/).
I'm getting API errors. Double-check that your Groq API key is correct and has not exceeded its rate limit.
The frontend shows a blank screen. Ensure the backend is running before opening the frontend. Check the browser console for errors.
npm install fails.
Make sure Node.js 18+ is installed. Run node --version to check.
Built for ML engineers who want to focus on ideas, not boilerplate.