Assistant Research Scientist (PREP0004904

Johns Hopkins, founded in 1876, is America's first research university and home to nine world-class academic divisions working together as one university.

Salary : $37,725-$56000 a year

Johns Hopkins University:
Whiting School of Engineering:
Office of Research and Translation

This position is part of the National Institute of Standards and Technology (NIST) Professional Research Experience Program (PREP). NIST recognizes that its research staff may want to collaborate with researchers at academic institutions on specific projects of mutual interest and, therefore, requires those institutions to be recipients of a PREP award. The PREP program involves staff from a wide range of backgrounds conducting scientific research across various fields.

Individuals in this position will perform technical work supporting the collaboration’s scientific research.

The PREP student will work with researchers in the Applied AI Research Group to develop a multi-agent consensus system that transforms AI-generated text into quantifiable scientific signals for antimicrobial peptide discovery. Antonio Cardone and Sarala Padi at NIST have developed STAMP (Species- and Topic-aware Representation Learning for AMP Discovery), a machine learning framework that predicts antimicrobial effectiveness. While STAMP generates explainability heatmaps indicating which peptide residues drive predictions, translating these model internals into actionable wet lab experiments requires high-confidence decision-making, as a single peptide synthesis requires 18-24 hours of laboratory resources.

This work addresses the challenge by engineering an ensemble of independent AI agents that analyze STAMP’s explainability heatmaps, generate structured “Hypothesis Cards,” undergo semantic consistency evaluation, and produce quantifiable consensus scores as decision gates for wet lab commitment.

Key responsibilities will include but are not limited to:

• Assist in building an ensemble of hypothesis generation agents using different large language model architectures (Claude, GPT-4, Llama variants)
• Assist in developing linguistic evaluation metrics to measure semantic consistency across agent outputs, including mechanism entailment, evidence overlap, and experiment design similarity
• Assist in creating parsers to transform STAMP’s Gradient × Input explainability heatmaps into structured hypothesis cards
• Implementing the end-to-end pipeline from STAMP heatmaps to hypothesis cards to consensus reports
• Benchmarking the system against historical experimental data
• Contributing to scientific manuscripts and technical documentation
• Developing production-ready code using Python and ML/AI frameworks with version control and testing best practices

§ Bachelor of Science degree in Computer Science (CS), or enrolled in a Master’s/Ph.D. program in CS, Computational Biology, Bioinformatics, or related field

§ Proficiency in Python programming

§ Familiarity with machine learning frameworks (PyTorch, Tensor Flow, or JAX)

§ Familiarity with natural language processing techniques

§ Version control (Git) and software engineering best practices

§ Understanding of deep learning architectures and statistical analysis

§ Willingness to work closely with wet lab researchers, biophysicists, and domain experts

§ Genuine interest in bridging AI/ML with biomedical applications

§ Strong communication skills for explaining computational concepts to non-CS audiences

§ Comfort working in interdisciplinary research environments

§ Self-motivated with strong problem-solving abilities

§ Strong oral and written communication skills.

If you are interested in applying for employment with Johns Hopkins University and require special assistance or accommodation during any part of the pre-employment process, please contact the HR Business Services Office  TTY users, call via Maryland Relay or dial 711. For more information about workplace accommodations at Johns Hopkins University for disabilities, medical conditions,…