Researcher, Model Architecture (UK) in City of London

About Cartesia

Our mission is to build the next generation of AI: ubiquitous, interactive intelligence that runs wherever you are. Today, not even the best models can continuously process and reason over a year-long stream of audio, video and text—1B text tokens, 10B audio tokens and 1T video tokens—let alone do this on-device. We’re pioneering the model architectures that will make this possible. Our founding team met as PhDs at the Stanford AI Lab, where we invented State Space Models or SSMs, a new primitive for training efficient, large-scale foundation models. Our team combines deep expertise in model innovation and systems engineering paired with a design-minded product engineering team to build and ship cutting edge models and experiences.

Role Responsibilities

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

What You’ll Bring

• Deep expertise in architecture design, with experience in researching or deploying advanced architectures (e.g., state space models, transformers, RNN variants, CNN variants).
• Strong understanding of how architectures interact with system constraints, including deployment in cloud environments or on-device.
• Proficiency in designing architectures that balance quality, efficiency, and adaptability across different use cases and modalities (e.g., vision, audio, text).
• Familiarity with generative modeling paradigms like autoregressive and diffusion models, and designing capabilities such as statefulness and conditioning in deep learning models.
• A proven research track record in top-tier ML/AI venues (e.g., NeurIPS, ICML, ICLR, CVPR) or demonstrable contributions to state-of-the-art architectures.
• Exceptional analytical and problem-solving skills, with a focus on experimentation and iterative refinement.
• Strong programming skills in deep learning frameworks such as PyTorch or TensorFlow, and experience with profiling tools for understanding model performance.

Nice-to-Haves

• Prior research or publications in state space models, efficient Transformers or other alternative architectures.
• Research or practical experience in designing architectures for multi-modal systems.
• Early-stage startup experience or a track record of rapid innovation in R&D environments.

Our culture

• We’re an in-person team based out of San Francisco, Bangalore & London. We love being in the office, hanging out together and learning from each other every day.
• We ship fast. All of our work is novel and cutting edge, and execution speed is paramount. We have a high bar, and we don’t sacrifice quality and design along the way.
• We support each other. We have an open and inclusive culture that’s focused on giving everyone the resources they need to succeed.

Our perks

• Lunch, dinner and snacks at the office.
• Fully covered medical, dental, and vision insurance for employees.
• Pension Plan.
• Relocation and immigration support.
• Your own personal Yoshi.

Seniority level: Mid-Senior level

Employment type: Full-time

Job function: Research, Analyst, and Information Technology

Industries: Software Development