Microsoft Azure AI Engineer (m/w/d)

Build AI/ML Solutions: Develop and implement AI/ML models to solve specific business problems. This could range from training predictive models (for example, forecasting equipment failures in manufacturing or customer behavior in insurance) to creating NL solutions like document classifiers or chatbots. You will use Azure Machine Learning for experiment tracking and model management, and write code (Python/SQL) for data processing and model training.Azure Cloud Implementation: Deploy AI solutions on Azure cloud infrastructure. Set up necessary resources such as Azure ML workspaces, Azure Functions or AKS for hosting models, and Azure Data Factory pipelines for data movement. Ensure that the deployment is done following Devoteam’s and Microsoft’s guidelines for security and compliance (important for enterprise clients in regulated industries like pharma and finance).Data Preparation & Feature Engineering: Work with data engineers to gather and prepare datasets required for machine learning. You will contribute to data preprocessing steps, writing data transformation scripts, defining features, handling data quality issues, to ensure models are trained on high-quality data. Utilize PySpark and SparkSQL in Azure Databricks or Microsoft Fabric for large-scale data processing when needed.Experimentation & Model Tuning: Conduct rigorous experiments to improve model performance. This involves trying out different algorithms or model architectures, tuning hyperparameters, and evaluating results. You’ll leverage Azure ML pipelines to automate training runs and compare metrics. For instance, you might compare a classical machine learning approach with a deep learning approach for a given problem and choose the best-performing model.Integration of AI Services: Make use of Azure’s AI services to accelerate development. Incorporate pre-built AI capabilities via Azure Cognitive Services (e.g. language translation, OCR, sentiment analysis) when appropriate instead of building from scratch. For example, if a project requires extracting text from PDFs and analyzing sentiment, you might use Azure Form Recognizer and Text Analytics as part of the solution before applying a custom model.MLOps & Monitoring: Implement the basics of MLOps for the solutions you build. Register models, create release pipelines for deploying them, and set up monitoring (for both system performance and model accuracy drift over time). Ensure that logging and alerting are in place so that any issues in production (like an API downtime or data drift causing model degradation) can be quickly identified and addressed.Collaboration & Documentation: Collaborate within a crossfunctional team, taking guidance from the Senior AI Engineer and Architect while also providing input based on your own expertise. Participate in design discussions, sprint planning, and code reviews. Additionally, document your work (datasets used, model assumptions, API specs) thoroughly to aid maintainability and knowledge transfer within the team and to clients’ IT teams.