• Independent literature research of academic texts and synthesis of findings • Identification of Machine Learning use cases in production environments • Conducting case studies and expert interviews • Documenting the use cases with use case diagrams

• Literature and market research on current trends and technologies in MLOps and their application in maintenance • Analysis of specific maintenance requirements for AI/ML models and their operation • Development of a tailored MLOps approach that facilitates the integration of AI/ML into maintenance • Derivation of practical recommendations and preparation of practice-oriented documentation

• Design, build, and maintain training and inference pipelines for traditional ML and LLM-based systems • Develop predictive models using regression, time-series, and probabilistic techniques • Build and refine confidence scoring systems to ensure model reliability • Integrate LLMs (OpenAI, HuggingFace) across products using APIs, LangChain, fine-tuning, and RAG pipelines • Conduct bias, drift, error, and fairness analysis to maintain model transparency and robustness • Implement and manage vector databases to support retrieval-augmented generation pipelines • Collaborate closely with engineering and product teams to bring ML-powered features into production environments • Apply MLOps principles: model monitoring, retraining workflows, versioning, and CI/CD for ML • Provide mentorship to team members and contribute to shaping the company’s AI/ML engineering standards • Stay current with industry advances and proactively recommend emerging tools, frameworks, and best practices

• Lead conceptual designs and create frameworks to predict a variety of outcomes in different scenarios • Build, deploy, and optimize machine learning models • Collaborate with other data scientists and stakeholders on projects • Develop solutions in Python • Develop production-grade solutions • Work in Hadoop, Redshift, and Spark • Translate business and product questions into analytics projects • Communicate clearly over written and oral channels