Senior Machine Learning Engineer, Performance

About us

Scribe is where exceptional people come to do the best work of their careers. More than 90% of the Fortune 500 use Scribe to automatically create step-by-step guides and streamline knowledge sharing. We’re growing fast — since our founding in 2019, we’ve grown to over 2.5 million users across 450,000 businesses. Based in San Francisco, we’ve raised $55M in funding from top-tier investors and are honored to have been named Fortune’s Next Billion Dollar Startup in 2024. Join us in our mission to unleash and up-level the world’s know-how!

How we work

We are builders aspiring to master our crafts. We care deeply about our teammates and want to win, together. We embrace the following values:

• Accelerate impact
• Raise the bar
• Make our users heroes
• Clear is kind
• Rapid learning machine
• One team one dream

About this Role

Scribe is a productivity automation company based in San Francisco. We are seeking a highly motivated and skilled Machine Learning / Applied Research Engineer to join our team. You will work on cutting-edge research projects to build the future of agents (see our work on ScribeAgent here) and rapidly develop, test, and deploy groundbreaking AI-powered software to our millions of users. We are constantly pushing the envelope with cutting-edge results and are now looking for more talent to come join us and delight our millions of users.

You can expect to

• Explore cutting-edge techniques in the artificial intelligence field and translate these innovations into valuable features for our millions of users.
• Design and implement efficient attention mechanisms and memory management strategies to help boost model performance.
• Optimize large language model inference through CUDA kernel development and KV cache management (e.g., quantization, pruning).
• Collaborate with the engineering team to balance performance and model quality
• Create benchmarking infrastructure for optimization experiments.
• Design experiments and rapidly iterate to increase the performance of our models with regard to evaluation metrics you will help define.
• Collaborate with our very talented engineering and product management teams.

You could be a great fit if

• You have a relevant degree from a top ML program — preferably 2+ years of industry experience working deep in the weeds on hard ML problems.
• You have strong software engineering skills (including Python, Jupyter, etc.)
• You have expertise in CUDA programming and GPU architecture.
• You have a strong understanding of transformer architectures and attention mechanisms.
• You have expert-level PyTorch development skills.
• You have strong communication skills.
• You have a track record of successfully owning projects from start to finish.

Bonus

• You have startup experience (not required, but we build and move fast!)
• You have proven contribution to open-source projects or publications in machine learning, statistics, computer science or related technical fields.
• You have a deep understanding of systems engineering to build scalable solutions.

Full-Time US Employee Benefits Include

• Some of the nicest and smartest teammates you’ll ever work with
• Competitive salaries
• Comprehensive healthcare benefits
• Exciting and motivating equity
• Flexible PTO
• 401k
• Parental Leave
• Commuter Benefits (SF office employees)
• WFH Stipend

Compensation

$160-$200k USD base + equity + benefits. We consider several factors when determining compensation, including location, experience, and other job-related factors.

At Scribe, we celebrate our differences and are committed to creating a workplace where all employees feel supported and empowered to do their best work. We believe this benefits not only our employees but our product, customers, and community as well. Scribe is proud to be an Equal Opportunity and Affirmative Action Employer.