Machine Learning Engineer, Model Integrations
$170k - $240k • San Francisco Office • FullTime
Posted 1mo ago
About the job
Nunchux AI is building infrastructure to make multimodal generative AI faster, cheaper, and easier to build on. Founded by MIT PhDs and CMU Professor, the company leverages deep research expertise and production systems experience. The Role focuses on bringing new image and video models to the Modelverse platform, prioritizing speed to launch. You will own the initial integration, coordinate deployment with the cloud team, and work with product engineers on API and Modelverse integration. After launch, you will hand off optimization tasks and maintain the shared code for model integration and serving.
Responsibilities
- Prototype new image and video models quickly, running test examples and identifying launch requirements.
- Coordinate model deployment with the cloud team and API/Modelverse integration with product engineers for the first release.
- Automate manual steps and simplify handoffs to speed up the launch process.
- Build reusable adapters and launch scripts to reduce the effort for integrating new models.
- Verify end-to-end integration, including outputs and error handling, and document known limitations.
- Fix bugs and add support for new model interfaces, providers, and modalities in the shared integration and serving code.
Requirements
- Experience running and adapting image or video generation models, or integrating provider APIs.
- Strong proficiency in Python and PyTorch, including reading model code, adapting inference pipelines, and debugging.
- Experience deploying models or building API integrations, including working with serving tools, reading logs, and debugging failed requests.
- Ability to quickly learn unfamiliar model code and get prototypes working.
- Capacity to keep the first release focused, resolve launch blockers, and coordinate with teammates.
Benefits
- Equity
- Health insurance
- 401(k)