Overview

Are you passionate about advancing deep learning research and developing general human-like machine intelligence? Join us as a Software Engineer, Machine Learning. In this role, you will collaborate closely with a senior member of our research team to work on cutting-edge deep learning projects, infrastructure, and tooling.

Responsibilities

Research Implementation: Read and implement various research papers to improve architecture efficiency, training speed, loss metrics, and fine-tuning methods.

Experimental Design: Devise and conduct scientifically rigorous experiments to validate the effectiveness of proposed adjustments for solving Imbue’s target problems using LLMs.

Fundamental Understanding: Enhance our knowledge of LLM fundamentals, focusing on areas such as chain-of-thought, reasoning, generalization, hallucination, and grokking.

Hyperparameter Optimization: Utilize our existing framework to perform hyperparameter sweeps, and contribute to the development of new features for the sweep code.

Infrastructure Collaboration: Work with infrastructure engineers to develop monitoring systems, informative logging, and maintenance guides for long-running experiments.

Dataset Improvement: Collaborate with data and product engineers to identify and implement improvements in our training and evaluation datasets.

Model Performance: Conduct scientific research and development to create LLM models that perform effectively in practical applications.

Example Projects

Implement improvements from recent research papers to enhance model architecture and training methods.

Design and run experiments to test the impact of new techniques on LLM performance.

Investigate and optimize aspects of LLM behavior such as reasoning and generalization.

Perform hyperparameter sweeps and contribute to the development of new features in our hyperparameter tuning framework.

Develop robust monitoring and logging systems for long-running experiments in collaboration with infrastructure engineers.

Improve the quality and effectiveness of training and evaluation datasets in partnership with data and product engineers.

Requirements

Requirements

Python Proficiency: Highly comfortable writing Python code.

PyTorch Experience: Familiar with PyTorch and experienced in training deep neural networks.

Open Source Enthusiasm: Excited to work on and contribute to open-source projects.

Best Practices: Passionate about engineering best practices and maintaining high-quality code standards.

Self-Direction: Able to work independently and manage your own tasks effectively.

Execution Excellence: Demonstrated ability to get things done efficiently and effectively.

Benefits

Benefits

Competitive Salary: $175,000 - $325,000 annually.

Health Insurance: Comprehensive medical, dental, and vision coverage.

Retirement Plans: 401(k) plan with company matching.

Paid Time Off: Generous PTO policy including vacation, sick leave, and holidays.

Professional Development: Opportunities for continuous learning and career growth, including access to conferences, workshops, and online courses.

Flexible Work Arrangements: Options for remote work and flexible scheduling to support work-life balance.

Parental Leave: Paid parental leave for new parents.

Wellness Programs: Access to mental health resources, wellness programs, and fitness reimbursements.

Employee Assistance Program: Support for personal and professional issues through our EAP.

Stock Options: Equity options to share in the company’s success.

Commuter Benefits: Pre-tax commuter benefits for public transportation and parking.

Technology Stipend: Annual stipend for tech equipment and home office setup.

Company Events: Regular team-building activities, social events, and company retreats.

Diversity and Inclusion: Commitment to fostering an inclusive and diverse workplace.