ABOUT TENNR:

Today, when you go to your doctor and need to be referred to a specialist (e.g., for sleep apnea), your doctor sends a fax (yes, in 2024, 90% of provider-provider communication is a 1980s fax). These are often converted into 20+ page PDFs, with handwritten (doctor’s handwriting!) notes, in thousands of different formats. The problem is so complex that a person has to read it, type it up, and manually enter your information. Tennr built RaeLLM (7B—trained on 3M+ documents) to read these docs, talk to your doc to ensure nothing is missed, and text you to help schedule your appointment so you can get better, faster.
Tennr is a NYC-based tech company that launched out of Y-Combinator and is backed by Lightspeed Venture Partners, Andreessen Horowitz, Foundation Capital, The New Normal Fund, and other top investors.

QUALIFICATIONS

• 3+ years of experience (post BS/MS) in an ML research/engineering role
• Proven track record of building and maintaining scalable web applications, particularly in high-volume workflow automation and data processing.
• Experience integrating machine learning models into production environments
• Can efficiently translate open-ended problems into actionable solutions
• Familiarity implementing novel NLP research ideas and techniques. Prior publications in top conference journals is a plus.
• Prior experience in a startup environment is a plus.

Machine Learning Engineers at Tennr are expected to wear a variety of hats. In the role, you will be expected to do the following:

• End-to-end product development: architect, train, deploy, and monitor models that drive direct customer value across our product.
• Data processing and ML Ops: optimize scalable data processing pipelines in our platform and maintain machine learning infrastructure.
• Backend integration: design and maintain complex workflows that leverage machine learning to drive automation.
• Product evaluation: Collaborate with sales and customer success teams to respond to feedback from our customers and prospects.
• Custom models: Fine-tune LLMs and VLMs for medical document understanding tasks