Scientific Software Engineer | Kumar Lab (Machine Learning)

Location: Bar Harbor, Maine, United States

Salary: $85,987–$143,962 annually, determined by years of experience.

Company

hirify.global is an independent, nonprofit biomedical research institution focused on genetics, disease research, therapeutic discovery, and scientific tools for the biomedical community.

What you will do

• Design, train, evaluate, and deploy computer vision and machine learning models for quantitative analysis of mouse behavior.
• Build production pipelines for continuous, petabyte-scale video processing and near-real-time inference.
• Develop and harden algorithms for segmentation, pose estimation, instance assignment, identity tracking, action recognition, and behavior localization.
• Evaluate model performance across strains, experimental rigs, sites, environments, and genetic backgrounds.
• Optimize models for edge inference, including quantization and runtime performance, and apply robust MLOps practices.
• Collaborate with scientists and project sponsors, review code, document systems, mentor trainees, and contribute to publications and open-source releases.

Requirements

• Bachelor’s degree required; a master’s degree in computer science, machine learning, or a related field is preferred.
• Approximately three years of hands-on experience required; five years preferred.
• Strong machine learning experience, including the ability to select, implement, evaluate, and explain methods and architectures.
• Production experience with PyTorch, training and evaluation infrastructure, model versioning, deployment, and data pipelines.
• Working familiarity with Python, ffmpeg, C++, Cython, SQLite, PostgreSQL, SLURM, Bash, Go, containers, Kubernetes, object storage, and cloud technologies such as GCP and AWS.
• Strong communication, prioritization, documentation, and collaboration skills, with interest in biological sciences or willingness to learn the domain.

Nice to have

• Experience with biological sciences, video analysis, animal behavior, or real-world computer vision data.
• Experience using LLM-based coding assistants such as GitHub Copilot or Claude Code while maintaining code quality and correctness.
• Experience with model generalization, continual learning, edge-case mining, and efficient deployment.

Culture & Benefits

• Work in a nonprofit biomedical research environment focused on therapeutic and mechanistic discovery.
• Collaborate with biologists, institutional stakeholders, external collaborators, and computational researchers.
• Contribute to scientific publications, open-source releases, and model releases.
• Support the development of digital measures for neurodevelopmental and neuropsychiatric disorders.

Hiring process

• Submit a CV and a paragraph describing the problem of interest and one relevant project with a link.
• For video or behavior-related work, explain the train/test split and the reasoning behind it.