Cartesian Systems
Machine Learning Engineer
About this role
Cartesian seeks a Senior Machine Learning Engineer to own end-to-end ML problems for their indoor positioning and item localization platform deployed across retail stores globally. You'll balance research and production, working across wireless signals, time series, and spatial reasoning to drive impact at a fast-growing MIT spinout.
What you'll do
- Design, develop, deploy, and monitor deep learning models for indoor positioning and item localization end-to-end
- Build and improve training, inference, and evaluation pipelines from prototype to production
- Optimize models for accuracy, coverage, latency, and cost while managing trade-offs
- Develop benchmarking tools and datasets for real-world, at-scale performance evaluation
- Collaborate with engineering and product teams to prioritize features and ship to enterprise customers
- Improve team engineering standards through reusable infrastructure and technical leadership
What they're looking for
- Machine learning (supervised/unsupervised learning, embeddings, metrics, evaluation)
- Python and PyTorch
- Production software engineering and code quality
- Data analysis and validation
- Time-series modeling and signal processing
- Computer vision or multi-sensor fusion
- Model optimization and deployment
- Cross-functional collaboration
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
Cartesian Systems
Cartesian Systems builds an indoor spatial intelligence platform for retail that uses machine learning to enable positioning and perception capabilities. The company is hiring ML Scientists and engineers to develop and deploy perception models and features for enterprise retail customers.
View all jobs at Cartesian SystemsLikely interview questions
- Walk us through a machine learning system you took from an ambiguous problem definition to production—what were the key challenges and trade-offs?
- How do you approach debugging when a model performs well in evaluation but fails in the real world at scale?