You will be an integral part of Data Sciences at PinkDx, building and implementing a machine learning/AI/statistical modeling engine that develops classification models using a comprehensive menu of genomic features generated from next-generation sequencing (NGS) platforms, clinical features, and adjudicated truth labels. You will generate candidate models and evaluate their performance using cross-validation, iterative training, model-building and perform lock and final validation. You are proficient in state-of-the-art machine learning/AI/statistical modeling methods and will evaluate their suitability for a wide variety of genomic and non-genomic datasets.
Essential Functions:
Utilize internal data analysis pipelines to support R&D objectives on aggressive timelines.
Supervise the building of an end-to-end machine learning pipeline that builds, evaluates, iterates, and tests performance of increasingly improved and stable classifiers on relevant biological and clinical datasets.
Implement feature engineering processes to identify, extract, evaluate, and incorporate a comprehensive suite of genomic features derived from both RNA and DNA whole-genome assays into an automated pipeline that supports rapid generation of machine learning models suitable for subsequent evaluation.
Collaborate with wet-lab scientists to perform end-to-end analysis of experiments including design, processing, QC, analysis, and interpretation of results.
Communicate complex multifactorial relationships and insights derived from biological data via compelling visualization using state-of-the-art best practices.
Identify and utilize external genomic data sources to implement and evaluate algorithms described in published literature with a goal towards improving the current state of diagnosis.
Work independently on complex projects, anticipate, and overcome challenges, and negotiate requirements and timelines to deliver results that support company objectives.
Supervise the work of junior colleagues when appropriate.
Work cross-functionally with internal teams and external scientific advisors to develop and implement analysis solutions and regularly present and participate in cross-functional discussions.
Adhere to timelines, design controls & best practices for data analysis.
Stay current with new machine-learning tools and software technologies.
Contribute positively to the building of culture at a fast-paced start-up.
Education & Experience:
PhD in Statistics, Mathematics, Bioinformatics, Genetics, Computer Science or a related field, with a focus on Machine Learning required.
8+ years of algorithm development and supervised and unsupervised machine learning/AI experience required.
5+ years of experience supporting diagnostics product development is preferred.
Expert in one of these languages is required: R, Python, C#, or C++
Experience building classifiers with underdetermined systems preferred.
Experience in NGS, RNA-Seq, whole transcriptome analysis is required.
A background in oncology is a plus.
Experience with workflow managers, container solutions, job schedulers, and cloud computing (AWS, Azure, Google cloud) is a plus.
Additional Skills:
Attention to detail and ability to prioritize tasks to meet critical deadlines.
Excellent verbal and written communication skills, to both expert and lay team members.
Able to collaborate effectively with the study team, cross-functional team members, and external partners.
Strong proficiency in Microsoft Office 365 and related products.