Cartesia
Researcher, Models
$180K–$350K
*HQ - San Francisco, CA · On-site · Full-time · Posted Dec 12, 2024
- Visa sponsorship
- Visa sponsorship support is assessed case by case and depends on the role and location; success is not guaranteed.
About the job
Research and develop neural network architectures that improve model quality, efficiency, adaptability, and deployment in real-world systems.
What you'll do
- Design and evaluate novel architectures, including state space models, efficient transformers, and hybrid architectures.
- Investigate capabilities such as statefulness, long-range memory, and conditioning, and study architecture trade-offs in scalability, robustness, latency, and energy efficiency.
- Build evaluation tools and collaborate with cross-functional teams to translate research into scalable systems.
What they ask for
- Expertise researching or deploying advanced architectures, such as state space models, transformers, RNN variants, or CNN variants.
- Understanding of architecture interactions with system constraints, including cloud or on-device deployment.
- Proven research track record at top-tier ML/AI venues or demonstrable contributions to state-of-the-art architectures.
- Strong programming skills in PyTorch or TensorFlow, and experience with profiling tools.
Nice to have
- Research or publications in state space models, efficient Transformers, or other alternative architectures.
- Research or practical experience designing architectures for multi-modal systems; early-stage startup experience or rapid R&D innovation.
Pay
$180K – $350K (base salary)
Tools
state space models, SSMs, transformers, RNN variants, CNN variants, PyTorch, TensorFlow
AI focus
Research and develop neural network architectures for AI models, including capabilities such as statefulness, long-range memory, and conditioning.
Model-written summary · the original posting is authoritative