Machine Learning Engineer

University of Illinois - Cambridge, MA

Full-time - Mid Level

Hybrid - Cambridge, MA

Educational Services

About the position

The Center for Computational Biomedicine (CCB) at Harvard Medical School is seeking a Machine Learning Engineer to develop large language models (LLMs) aimed at enhancing medical education and clinical decision-making. This role is crucial in leveraging computational techniques to improve health outcomes and advance scientific discovery through collaboration with a multi-disciplinary team. The engineer will contribute to the center's mission by creating medical LLMs that support educational initiatives and clinical processes.

Responsibilities

• Develop, implement, and optimize medical large language models tailored to the needs of medical education and clinical decision support.
• Collaborate with interdisciplinary teams comprising biologists, clinicians, and data scientists to understand domain-specific requirements and translate them into computational solutions.
• Stay updated with the latest advancements in deep learning and machine learning to ensure the models developed are state-of-the-art.
• Develop infrastructures for data transformation and ingestion.
• Build AI models that make predictions based on large quantities of data.
• Explain the usefulness of the AI models created to stakeholders.
• Transform machine learning models into APIs to interact with other applications.

Requirements

• A Master's or PhD in Computer Science, Computational Biology, or a related field.
• Minimum of seven years' post-secondary education or relevant work experience, education will count towards experience.
• Minimum of 3 years of hands-on experience in developing complex deep learning solutions to tackle scientific challenges.

Nice-to-haves

• Proficiency with the Python deep learning software stack, particularly expertise in PyTorch, Numpy, and related packages.
• Experience handling and processing large and diverse datasets, especially medical texts, journals, or electronic health records.
• Ability to collaborate effectively with non-technical stakeholders, such as doctors and medical researchers.
• Experience with experiment tracking and project management tools, notably frameworks like Weights & Biases.
• Prior experience in fine-tuning large language models for specific tasks.
• Demonstrated experience in optimizing deep learning models for better performance and efficiency.
• Understanding of biology and/or medicine to bridge the gap between pure machine learning and its applications in the medical field.
• A track record of publications in technical conferences or journals.

Benefits

• Paid Time Off: 3-4 weeks of accrued vacation time per year, 12 accrued sick days per year, 12.5 holidays plus a Winter Recess, 3 personal days per year, and up to 12 weeks of paid leave for new parents who are primary caregivers.
• Comprehensive medical, dental, and vision benefits, disability and life insurance programs, along with voluntary benefits.
• Child and elder/adult care resources including on-campus childcare centers, Employee Assistance Program, and wellness programs.
• University-funded retirement plan with contributions from 5% to 15% of eligible compensation, based on age and earnings with full vesting after 3 years of service.
• Tuition Assistance Program including $40 per class at the Harvard Extension School and reduced tuition through other participating Harvard graduate schools.
• Tuition Reimbursement Program that provides 75% to 90% reimbursement up to $5,250 per calendar year for eligible courses taken at other accredited institutions.
• Professional Development programs and classes at little or no cost, including through the Harvard Center for Workplace Development and LinkedIn Learning.
• Various commuter options including discounted parking, half-priced public transportation passes, and biking benefits.
• Access to Harvard athletic and fitness facilities, libraries, campus events, credit union, and discounts to various services and cultural activities throughout metro-Boston.