Machine Learning Research Engineer (m/f)

Our R&D department conducts applied AI research in two core domains: medicine and data privacy. In healthcare, we develop AI models for early cancer detection, treatment response prediction, and clinical decision support, working with European research partners and real clinical datasets. In data privacy, we are building AutoSynth: a synthetic data generation platform for tabular data, with a strong emphasis on medical use cases enabling organizations to share and work with sensitive data without compromising patient privacy. Every research initiative at Atomic Intelligence must lead to one of two clear outcomes: either publishable results that advance the state of the art through conference or journal papers, or a production-ready feature that becomes part of one of our digital products. There is no research without delivery. In this role, you will split your time between research projects and product work. Some weeks you'll be reading papers and experimenting with new architectures. Other weeks, you'll be debugging a training pipeline or preparing data for a new ML model. This role offers the opportunity to grow across the full ML lifecycle, from data preparation and model design to published research and production deployment. Key Responsibilities Research & Innovation: - Design and implement new layers and components for deep learning models - Explore fine-tuning approaches using reinforcement learning and other advanced methods - Conduct research in domain-specific AI applications: medical imaging, clinical tabular data, synthetic data generation, audio separation, and NLP - Develop and evaluate generative models (VAE, GAN, diffusion) for tabular synthetic datawith a focus on medical and privacy-sensitive use cases - Experiment with hybrid architectures that combine different ML approaches - Publish findings and contribute to the scientific community through papers and open-source work Technical Implementation: - Design and validate model architectures through experimentation - Implement custom loss functions, optimization strategies, and training pipelines - Turn research breakthroughs into efficient, production-ready implementations that become core features of our products - Collect, prepare, and preprocess data for model training Collaboration & Leadership: - Contribute to research planning and project roadmaps - Work with European research institutions on ongoing funded projects - Stay current with emerging AI research and bring relevant ideas to the team

- 3+ years of professional experience in ML research or production environments - Degree in Computer Science, Mathematics, Physics, or related fields (PMF, FER, FOI, or similar) - Strong foundation in mathematics, algorithms, and data structures - Proficient in Python and experienced with ML frameworks (PyTorch preferred, TensorFlow acceptable) - Ability to read, understand, and improve upon academic papers - Experience with model architecture design and experimental methodology - Experience with experimental methodology: ablation studies, hyperparameter tuning, reproducibility - Experience with at least one of the following domains: medical data (imaging or tabular clinical data), generative models, synthetic data - Analytical and problem-solving skills, with the ability to translate research into technical solutions Preferred Qualifications - Experience with modern AI systems: RAG architectures, intelligent agents, LLM orchestration - Track record of published papers or open-source contributions - Experience with reinforcement learning or advanced fine-tuning techniques - Familiarity with distributed training and large-scale data processing - Experience with containerization (Docker) and ML deployment workflows