The role, as published
About the company
Our client is a fast-growing software company.
The role
Raydar is recruiting for this role on behalf of our client. Lead the development of large multimodal machine learning models that drive physical hardware, covering data preparation, training, evaluation and real-world testing. This is a hands-on, applied position centered on improving how quickly and dependably models perform once deployed.
What you'll do
- Train and refine deep learning models so they meet reliability and latency goals for a live customer deployment.
- Test different datasets, training approaches and hyperparameters to learn which changes lift success rates.
- Design and grow compute infrastructure that lets many training runs happen in parallel across GPU resources.
- Take responsibility for the whole model lifecycle, from assembling data through training, evaluation and testing on real machines.
- Increase experimentation throughput across multiple workstreams as the business grows.
What we're looking for
- 6+ years of experience in machine learning engineering.
- A history of taking trained models into production on physical devices.
- Background running large-scale training jobs spread over several GPUs.
- Familiarity with multimodal or foundation models applied to robotics.
- Practical skill training neural networks in PyTorch or JAX.
- Knowledge of transformer architectures and generative modeling techniques.
- Exposure to policy learning methods such as learning from demonstrations.
- A bachelor's degree in computer science, machine learning, robotics, electrical engineering or a related field, or equivalent practical experience.
- Self-directed approach to unclear, open-ended problems, with a record of producing functioning solutions.
- U.S. citizenship is required.
Location and work model
- New York, NY, United States
- On-site, 5 days per week in the office
- Full-time
Compensation, as the posting describes it
- Equity