Requirements
- Over four years of industry experience in machine learning and deep learning, especially in health sciences or regulated sectors, with demonstrated success in deploying algorithms to production.
- Experience with digital signal processing and statistics, with a focus on selecting appropriate tools, which may not always involve machine learning or deep learning.
- Proficiency in PyTorch or similar deep learning frameworks for training, developing, and deploying models.
- Knowledge of recent advances in deep learning, such as Transformer architectures, large-scale modeling, and extensive model training.
- Adherence to best practices in software and ML engineering, including testing, version control, code reviews, documentation, containerization, CI/CD, and experiment tracking.
- Familiarity with biosignals, medical imaging, or large time-series datasets, or a strong interest in learning about these areas.
- Thrives in team settings, valuing collaboration, open communication, and ongoing feedback for shared achievements.
- Ability to simplify, discuss, and present complex technical subjects appropriately for various audiences, both internal and external.
- Enthusiasm for engaging in the full algorithm development lifecycle, from scoping and data preparation to development, validation, documentation, deployment, and client interaction.
Nice to Have
- Proficiency in PyTorch is preferred.
Benefits
- Equity
- Paid time off
Work Arrangement
Hybrid — Boston, New York City, Paris
What success looks like
- Participate in and lead the entire biosignal-based algorithm development lifecycle for medical devices including specifications and requirements gathering, data curation and labeling, development, failure-analysis, production, maintenance, and documentation.
- Select, implement, and develop the most appropriate method for each problem, knowing when to apply deep learning techniques and when other methods are more effective.
- Enhance our internal deep learning and machine learning tools to boost team efficiency, introduce new model architectures and algorithmic techniques, and refine the codebase to encourage reusability where needed to enable rapid experimentation.
- Spread and improve our best practices to ensure algorithm implementations are user-friendly, well-documented, and thoroughly tested, including unit tests, comprehensive documentation, CI, and non-regression testing.
- Present results to key stakeholders and assist them in utilizing algorithms for client engagement.
- Support the client-facing projects to understand and shape the impact Beacon algorithms have for our customers, both for existing deployed algorithms, and future algorithm development.
What you will bring
- You have more than 4 years of industry experience in machine learning and deep learning, particularly in health sciences or other regulated fields, with a proven track record of bringing algorithms into production.
- You are experienced with digital signal processing (DSP) and statistics and care about using the right tool for the job, which in many cases might not be machine learning or deep learning.
- You are proficient in using PyTorch (preferred) or other deep learning frameworks for training, developing, and deploying deep learning models.
- You are familiar with latest Deep Learning advances (Transformer/ViT, large scale modeling, large model training, ...)
- You follow and adopt best practices in software and ML engineering, including testing, version control, code reviews, documentation, Dockerization, CI/CD, and experiment tracking.
- You are familiar with biosignals, medical imaging data, or large time-series datasets, or are enthusiastic about learning more in the domain.
- You thrive in a team environment, recognizing that collaboration, open communication, and continuous feedback are essential for collective success.
- You are able to distill, discuss, and present complex technical topics in a way that is appropriate for the audience at hand, both internally and externally.
- You are excited to participate in the entire algorithm development lifecycle, which spans scoping, data wrangling, algorithm development/experimentation, formal validation, quality/regulatory documentation, production deployment, and working with clients who might benefit from these algorithms.
