Researcher, Multimodal AI Architect

Cartesia; United States; Hybrid; USD 159,000 - 225,000; Full time

Benefits offered by this job

Compensation

Commuter allowance

Flexible PTO

Meals & Snacks

Equity package

Job summary

Cartesia is seeking an accomplished AI architect to push real-time multimodal intelligence. You will design and evaluate novel architectures, from state space models to efficient Transformers, balancing quality, efficiency, and deployment constraints across cloud and edge.

Join a fast-moving, research-driven team with strong publication credentials and a track record of impactful systems. London-based office with visa sponsorship as needed.

Qualifications

• Deep expertise in architecture design with research/deployment of advanced architectures such as state space models, transformers, RNNs or CNNs.
• Strong understanding of deploying architectures in cloud and on-device environments.
• Ability to balance model quality, efficiency and adaptability across multiple modalities and use cases.
• Proven research track record in top ML/AI venues or demonstrable state-of-the-art contributions.

Responsibilities

• Conduct groundbreaking research in neural network architecture design to advance state-of-the-art in alternative architectures (e.g., state space models, efficient Transformers, hybrid architectures).
• Design novel architectures that improve model quality, inference efficiency, and adaptability across deployment environments (cloud to device).
• Explore and develop capabilities like statefulness, long-range memory, and conditioning mechanisms for generalization.
• Investigate how architectural decisions impact trade-offs such as scalability, robustness, latency, and energy efficiency.
• Develop new frameworks/tools to evaluate architectural innovations and benchmark performance across research and production.
• Collaborate with cross-functional teams to translate architectural research into scalable, real-world systems.

Skills

State space models

Transformers

RNN/CNN variants

Python (PyTorch/TensorFlow)

Profiling & performance tuning

Tools

PyTorch

TensorFlow

Job description

Cartesia is seeking an accomplished AI architect to push real-time multimodal intelligence. You will design and evaluate novel architectures, from state space models to efficient Transformers, balancing quality, efficiency, and deployment constraints across cloud and edge.

Join a fast-moving, research-driven team with strong publication credentials and a track record of impactful systems. London-based office with visa sponsorship as needed.