Principal Vector Data Engineer

Johnson and Johnson Innovative Medicine (J&J IM), a pharmaceutical company of Johnson & Johnson is recruiting for a Vector Data Engineer.  This position has a primary location of Barcelona, Spain. The secondary location is Madrid. This is a hybrid role. 

Our expertise in Innovative Medicine is informed and inspired by patients, whose insights fuel our science-based advancements. Visionaries like you work in teams that save lives by developing the medicines of tomorrow.  

Join us in developing treatments, finding cures, and pioneering the path from lab to life while championing patients every step of the way. Learn more at https://www.jnj.com/innovative-medicine   

Position Summary:  

The Vector Data Engineer sits at the intersection of machine learning, signal processing, and neuroscience, with a focus on neurodegenerative and neuropsychiatric disorders. The successful candidate will contribute to foundational model development for longitudinal data integration and disease progression prediction, helping uncover biomarkers and insights into conditions such as Alzheimer’s disease, Parkinson’s disease, and Major Depressive Disorder. 

Key Responsibilities:  

• Design and implement vector embedding models to unify diverse biomedical data modalities, including: 

• EEG, PPG, accelerometry, biosensor signals 

• Speech and physiological recordings 

• Medical imaging (MRI, PET) 

• Develop quality control protocols for mobile-captured images and transform pixel-based images into vector representations. 

• Adapt and extend custom embedding methods from academic and research settings to build scalable foundation models. 

• Collaborate with cross-functional teams to integrate multimodal data for machine learning and clinical insight generation. 

• Contribute to the identification of digital biomarkers and predictive patterns in neurological and psychiatric conditions. 

Qualifications: 

• MS/PhD in Computer Science, Electrical Engineering, Biomedical, or related field. 

• Minimum 3 years of experience in multimodal data modeling, machine learning, or biomedical signal processing. 

• Familiarity with large pre-trained multimodal models (e.g., CLIP, MedCLIP, Flamingo) and biomedical/time-series adaptations (BioBERT, TimeSformer). 

• Strong proficiency in Python, with hands-on experience using PyTorch, TensorFlow, Hugging Face Transformers, and scikit-learn. 

• Signal & Speech Processing: Experience with Librosa, SpeechBrain, or similar libraries for acoustic and temporal feature extraction. 

• Proficiency in NiBabel for MRI and PET imaging, and familiarity with FSL, FreeSurfer, and NiLearn for neuroimaging analysis. 

• Multimodal Fusion: Experience with CLIP-like architectures and contrastive/self-supervised learning for multimodal data integration. 

• Understanding of clinical trial data and longitudinal monitoring frameworks, as well as experience with large-scale dataset curation and embedding evaluation.  

Strategic Impact:  

• Digital health and clinical multi-modal data assets transformed into interoperable, vectorized embeddings supporting scientific AI applications. 

• AI can perform semantic queries and reasoning over governed datasets. 

• Vector database infrastructure scales efficiently and complies with governance and lineage standards. 

• Accelerated insight generation across discovery, translational, and clinical domains. 

#JRDDS  #JNJDataScience