Position Overview: 

The Principal Machine Learning Scientist II will advance deep learning methods for biologics discovery at AbbVie.  

This individual contributor role focuses on advancing sequence- and structure-based deep learning models for protein and antibody design with emphasis on adapting them to the constraints of designing therapeutic antibodies. The work centers on building models that improve early decision-making in molecule prioritization and guide molecule optimization, while balancing multiple properties, in order to reduce drug discovery timelines and increase the probability of success in patients. 

A major focus of the role is extending protein modeling to support the complexities of biotherapeutics, including complex multispecific formats, sequence diversity, conformational flexibility, and induced-fit behavior, which present important modeling challenges and opportunities for innovation. 

The role requires close collaboration with antibody engineers, computational physicists, and biologists to translate domain concepts into modeling decisions that will drive therapeutic impact. 

Responsibilities

• Advance sequence- and structure-conditioned generative modeling for biologics using protein language models, diffusion-based generative models, and graph- and geometry-based deep learning methods. 
• Develop modeling strategies that improve therapeutic relevance, support multi-objective optimization, and surface liabilities earlier in the discovery process. 
• Develop predictive modeling approaches for multispecific biologics that optimize chain-complexation states, linker architecture, and higher-order molecular geometry to identify designs with favorable efficacy and developability. 
• Adapt protein models to CDR H3 loop behavior, including high diversity, conformational flexibility, and induced-fit effects. 
• Build rigorous internal benchmarking frameworks, establish evaluation standards for external AI/ML platforms, and lead technical due diligence on partner capabilities. 
• Use interpretability methods to uncover missing biological concepts, guide targeted model improvements. 
• Partner with deployment engineers to deliver scalable, stable production models and model monitoring workflows.