Senior Machine Learning Engineer - MLOps

Key Skills

AWS, Continuous Integration & Continuous Delivery - CI/CD, Data science techniques, Deep Learning, Docker, Google Cloud Platform (GCP), Kubeflow, Kubernetes-K8s, Linux Operating system, Machine learning techniques, MLflow, MLOps tools, Python Programming, PyTorch, TensorFlow

Job Description

As a Senior MLOps Engineer you will be the owner of our development environment, helping our team of NLP engineers train and test our models with popular solutions on the market such as DVC and MLFlow. You will also have control on how our ML product works within our production environments, helping bring better code practices into our deployed solutions.

Together with your colleagues you will form a small, cosy and self-organising team. As a team you are responsible for building the infrastructure, implementing the algorithms and converting them to production ready models. Additionally, you will be part of a larger, cross-functional and dynamic team consisting of developers, designers, medical consultants and data engineers.

Medical data is an interesting problem space, both in complexity and in societal value. The texts are a domain of its own with extremely high information density. It gives a real kick when users let us know they were able to collect data that they were unable to collect without CTcue, such as finding people with a rare disease that might take years otherwise to diagnose or rapidly building quality datasets for COVID research.

Key responsibilities:

• Mature and improve our CI/CD tools and components, and maintain CI/CD pipelines across multiple environments
• Python code optimization, containerization, deployment, versioning, and monitoring of its quality
• Data science models testing, validation and tests automation
• Develop and deploy scalable tools and services on our sites to handle machine learning training and inference
• Identify and evaluate new technologies to improve performance, maintainability, and reliability of our deployed machine learning systems
• Design and implement future cloud solutions, build MLOps on cloud (Azure)
• Apply software engineering rigor and best practices to machine learning, including CI/CD, automation, etc.

Your profile

You are a professional who is passionate about improving healthcare and having a real impact. A background in Machine Learning is nice to have, but we will also value DevOps experience and a solid interest in AI. You have an appetite to learn and to develop yourself, an innate curiosity, and you can bring to light clarity around abstract and unclear problems. You feel comfortable working in a self-organising company where there is little hierarchy. You feel at home in maturing our ML ops solution and helping you fellow data scientist create better products.

What we’re looking for:

• Bachelor's degree in Computer Science or Software Engineering
• Substantial experience in Python
• Ability to understand tools used by data scientist and experience with software development and test automation
• Ability to design and implement scalable solutions and ability to build MLOps pipelines on (cloud) solutions
• Experience with MLOps Frameworks like Kubeflow, MLFlow, DVC/CML, etc., experience with CUDA, Docker and Kubernetes
• Strong software engineering skills (Python), good understanding of Linux, knowledge of Deep Learning frameworks (such as PyTorch, Tensorflow, etc.)
• Experience in using MS Azure (preferred) or GCP, AWS services

The team

Today we are used to having most information - anywhere in the world - right at our fingertips. It may therefore surprise you that this is not the case at all for the hospital world. When we (CTcue department) started building the Patient Finder search engine, we thought we could build a simple application that could automatically identify patients for clinical trials. However, in the process of developing this, we have gradually come to see that we are building something much more fundamental.

The problem in hospitals is that the data in the Electronic Health Record (EHR) is not suitable for performing automated analyses. About 70% of the data consists of text (unstructured data), and standard BI tools cannot handle that well. The result is that in practice manual searches are still the default.

We have developed a data platform in which the unstructured data is organized using machine learning / NLP and then stored together with the structured data from the EHR. The platform contains two innovative applications with which the EHR data can be searched and collected.

Our aim has always been to support doctors in their work using data technology and it is with this in mind that we continue to develop and expand our platform. It goes without saying that privacy legislation forms a guideline in this. Our search engine is currently used daily in more than 40 hospitals in the Netherlands and Belgium. Our ambition is to expand the footprint of our platform to other key care and research hospitals across EMEA over the next few years.

Those who join us become part of a recognized global leader still willing to challenge the status quo to improve patient care. In this business unit, you will have access to the most cutting-edge technology, the largest data sets, the best analytics tools and, in our opinion, some of the finest minds in the Healthcare industry.