Computational Scientist - Multimodal AI & Biomarker Discovery

Job Description

Job Description

Olmeda BioSciences is an NIH-funded biotechnology company developing molecular and computational approaches to understand and predict eye disease through aqueous humor liquid biopsy, ophthalmic imaging, multimodal clinical data, and AI.

We are building an Aqueous Humor Atlas that integrates molecular profiles from the eye with OCT and fundus imaging, clinical phenotypes, treatment history, and longitudinal outcomes.

The Role

We are seeking a highly motivated, hands-on Computational Scientist to help build this platform and lead computational work across our ophthalmology and biomarker programs.

A major initial focus will be our prospective diabetic macular edema (DME) program integrating aqueous humor molecular data, OCT imaging, clinical characteristics, treatment history, and treatment response.

A central scientific question is whether molecular states measured in aqueous humor can be linked to retinal imaging phenotypes, and conversely whether non-invasive imaging can help infer underlying molecular states.

What You Will Do

• Build reproducible computational pipelines for proteomic, genomic, imaging, and clinical datasets
• Analyze and integrate high-dimensional molecular and multimodal clinical data
• Develop computational approaches linking aqueous humor molecular profiles with OCT and fundus imaging
• Apply statistical learning, machine learning, and AI to biomarker discovery and treatment-response prediction
• Develop and evaluate imaging representations and multimodal models
• Integrate public single-cell, proteomic, genomic, and ophthalmic datasets with internally generated data
• Identify confounders, batch effects, data leakage, and appropriate validation strategies
• Translate computational findings into biologically interpretable hypotheses and potential biomarkers
• Work directly with scientists, clinicians, and external collaborators

What We Are Looking For

• PhD in computational biology, bioinformatics, biomedical engineering, computer science, biostatistics, genomics, or a related quantitative field
• Strong hands-on programming skills in Python and/or R
• Experience analyzing real biological, imaging, or clinical datasets
• Strong foundation in statistics and machine learning
• Experience in at least one of the following: genomics, transcriptomics, proteomics, single-cell analysis, medical imaging, computer vision, or multimodal data analysis
• Ability to independently move from raw data through QC, analysis, modeling, validation, and biological interpretation
• Understanding of model validation, overfitting, confounding, and data leakage
• Strong scientific curiosity and ability to learn new data modalities quickly

Experience with PyTorch or related deep-learning frameworks, OCT/medical imaging, multimodal AI, proteomics, single-cell analysis, or longitudinal clinical data is a plus.

What Matters Most

We are less focused on years of experience than on scientific thinking, technical depth, and the ability to work directly with data and code. This is a hands-on role for someone who enjoys taking a new dataset from raw data to biological insight and building new computational approaches along the way.

Location

San Diego, California. This is a full-time position, and we prefer candidates who can work regularly from our San Diego office.

Olmeda BioSciences is an NIH-funded biotechnology company developing molecular and computational approaches for eye disease using aqueous humor liquid biopsy, multimodal ophthalmic imaging, clinical data, and AI. Our programs span biomarker discovery, treatment-response prediction, and the development of a multimodal Aqueous Humor Atlas.

Company Description

Olmeda BioSciences is an NIH-funded biotechnology company developing molecular and computational approaches for eye disease using aqueous humor liquid biopsy, multimodal ophthalmic imaging, clinical data, and AI. Our programs span biomarker discovery, treatment-response prediction, and the development of a multimodal Aqueous Humor Atlas.