Software Engineer (m/w/div)

The German Pension Insurance Federation is the largest German pension insurance company and thus a cornerstone of social security in Germany. We accompany people all their lives – from start of work to retirement age. 23.3 million insured persons, 9.6 million pensioners and 3.5 million employers rely on us. We maintain a highly complex and diverse IT system landscape. The methods and techniques used in this environment, including the hardware and software techniques used, offer an interesting and challenging working environment. For our AI laboratory, at the Hohenzollerndamm, we are looking for a Software Engineer (m/w/div) Location: Berlin Application deadline: 13.07.2025 Employment: Full time (part time is possible in principle) Date of entry: Immediate Call number: S5-011-2025 Remuneration: E14 TV EntgO-DRV Activity The AI laboratory bundles AI expertise and AI talent in a central unit in the digital strategy and digital transformation office. As an expert team for AI and application case factory, we bring KI to the DRV Bund and support all specialist departments as a contact partner*in with our support and consulting offer. With the help of AI technologies, we deliver new solutions in a steadily growing interdisciplinary team and thus provide pioneering work for the numerous challenges of a federal authority with more than 26,000 employees.

Develop and implement robust software architectures and design APIs for system integration of ML components (Machine Learning) in products Build and maintain CI/CD pipelines and implement containerization (Docker, Kubernetes) for ML workloads on cloud platforms Optimize the efficiency of ML models, in particular for inference in production environments Responsibility for the deployment, monitoring and maintenance of ML models in productive environments Working with the selection and integration of LLMs (Large Language Models) and NLP technologies (Natural Language Processing) Ensure high code quality through tests, code reviews and documentation as well as active collaboration in agile, interdisciplinary teams Close interaction with Data Scientists for converting prototypes into high-performance, production-ready applications

Completed scientific higher education (master, diploma or similar degree) with a quantitative focus; ideally the study direction MINT (mathematics, computer science, natural science or technology) or the social and economic sciences At least three years of professional experience in software development, including at least two years in machine learning or AI-based systems Detectable multi-year project experiences with agile approach models (for example Scrum, Kanban) in the context of AI products Very good programming knowledge in Python, including relevant ML libraries (for example PyTorch, scikit-learn, pandas) and the ability to write clean, testable code Funded knowledge of software architecture, design batteries for scalable systems and practical experience with cloud platforms (AWS, Azure) Expertise in containerization (Docker, Kubernetes) and orchestration as well as experience with MLOps concepts and tools (for example MLflow, Kubeflow) Knowledge of developing APIs with frameworks such as FastAPI or Flask Basic understanding of NLP concepts and the functioning of Large Language Models (LLMs), as well as familiarity with vector databases and Retrieval-Augmented Generation (RAG) Proactive communication behaviour, strong cooperation readiness and networked thinking in agile cooperation complete your profile

We offer you an interesting and varied activity with all the advantages of a large public employer Possibility to help shape the digital transformation of the DRV Bund Various projects such as KIRA, in which an AI-based application is developed for our audit service Individual training opportunities, diverse training opportunities and participation in specialist conferences promote technical growth and Communities of Practice (CoP) support the continuous exchange of knowledge An excellent view from the 16th floor of the Silberturm over the roofs of Berlin at Hohenzollerndamm Compatibility of family and profession through flexible working time models and the possibility of work in the home office; at least two days of presence per week Family service: benefit@work for parents and nurses Further information In order to occupy the position, we will have discussions with applicants who are in the narrower selection. Part of the selection process will be a work sample in the form of a programming task in Python. The activity requires very good knowledge of German in word and writing. The German Pension Insurance Federation has set itself the goal of promoting women. We therefore look forward to receiving applications from women with particular interest. Persons with a severe disability or persons with equal status within the meaning of § 2 para. 2 and 3 SGB IX are preferably taken into account with the same suitability. We welcome applications from people of all nationalities. JOBV1_EN