Machine Learning Engineer
Glitch Devs
- Work mode
- Remote
- Job type
- Full time
- Team
- AI & Data
- Salary
- USD 900 to 1,600 / month
Posted 2 months ago
About the role
You will take machine learning work from a notebook to something running in production. That means preparing data, training and evaluating models, and then making them available through an API that other systems can call.
Our projects vary. Some use classic ML on tabular data, some use language models or embeddings. We care less about the newest paper than about whether the model works on real data and keeps working after launch.
The work:
- Prepare and clean datasets
- Train, tune and evaluate models, and report the results honestly
- Deploy models behind APIs
- Monitor accuracy and cost once they are live
- Explain results and limits to people who are not engineers
Examples of what you might work on:
- A model that predicts which customers are likely to cancel, with an honest report on where it fails
- A document classifier that sorts incoming files and flags the ones it is unsure about
- A service that serves a model through an API and tracks its accuracy after launch
Tools you will use:
Python, pandas, scikit-learn, PyTorch or TensorFlow, SQL, Docker, MLflow, a cloud provider such as AWS or GCP.
Your first months:
- First weeks: Study the current data and models, and reproduce an existing result.
- By month two: Improve one model or evaluation step, and report the change clearly.
- By month three: Take a model from notebook to a deployed service with monitoring.
Requirements
- 3 or more years working on machine learning in a real setting
- Strong Python, with pandas and scikit-learn, and PyTorch or TensorFlow
- Experience evaluating models properly, not just reporting accuracy
- Able to deploy a model as a service
- SQL
Nice to have:
- Experience with embeddings, vector search or fine-tuning language models
- MLflow or similar experiment tracking
- Cloud experience on AWS, GCP or Azure
Practice for this role
Try a free 20-question test built for this role. It takes about 20 minutes and you get an explanation for every answer. It is not part of the application.
Take the practice testKnow someone who’d be a good fit?
Similar roles
See all roles →Machine Learning Engineer
Apply now