About the role
Nouvells is looking for Machine Learning Engineers who can turn data and machine learning concepts into reliable, production ready systems. You'll contribute to developing, evaluating, optimizing, and deploying machine learning models across real-world applications.
As our Machine Learning Engineer, you will turn machine learning prototypes, representations, and foundation models into robust, scalable production services. You will take ownership of the model execution lifecycle: packaging weights, optimizing inference runtimes, eliminating CPU-GPU memory transfer bottlenecks, and building resilient APIs. You will collaborate directly with our Data Scientists, MLOps Engineers, and Backend Platform developers to ensure that models train reliably, deploy without regression, and serve predictions under high concurrency.
Key responsibilities
- Convert raw model checkpoints (PyTorch, Transformers) into optimized inference engines using ONNX Runtime, TensorRT, TorchScript, or OpenVINO.
- Profile production inference code to identify and eliminate CPU-GPU serialization bottlenecks and memory leaks.
- Build shadow-testing and canary-routing infrastructure to benchmark candidate models against live production traffic before full deployment.
- Write automated integration tests, model output regression tests, and performance load tests within CI/CD pipelines.
Qualifications
- Bachelor’s Degree and equivalent practical experience