Staff / Principal Research Scientist
Inworld AI develops AI products for growing applications, helping developers go from prototype to production faster. Their offerings include advanced text-to-speech (TTS) technology and an upcoming Runtime product, aimed at enhancing consumer applications with expressive, real-time voice AI.
Maintainer signals as of 8/23/2026
Funding history
About Inworld AI
Inworld develops AI products for consumer applications. They offer a text-to-speech model that aims for high quality with better pricing, lower latency, more control, local serving options, and open training code. They also have a product called Inworld Runtime, which is currently in private preview. Their services are used by partners like XBOX, Ubisoft, NVIDIA, and Meta. They focus on helping developers go from prototype to production faster and increase experimentation velocity to deploy new AI improvements daily.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
Own ambiguous research problems end-to-end, from framing questions and running experiments to shipping results. Work across foundation models, agentic systems, multimodal models, tool use, and test-time compute, with a focus on impact and evaluation.
Requirements
- Experience with foundation models: training, new architectures, RL, reward modeling, scaling
- Experience with evaluation: benchmarks, eval loops, quality measurement, LLM-as-judge, failure analysis
- Experience with frontier topics: multimodal models, agents, tool use, test-time compute, world models
- Published research at ICML, ICLR, NeurIPS, EMNLP, ACL, or AAAI
- PhD in ML/NLP or equivalent practical experience
- Public work such as non-trivial AI side projects, interdisciplinary experiments, or open-source contributions
- Full-stack research ownership experience
- Must already have the legal right to work in the United Kingdom
Responsibilities
- Research and develop state-of-the-art foundation models
- Optimize realtime inference
- Design evaluation loops and quality measurement systems
- Frame research questions, run experiments, and ship results
- Perform failure analysis on models and systems
- Work on frontier topics such as multimodal models, agents, tool use, test-time compute, and world models
Benefits
- Equity
