Staff / Principal Research Scientist

About Inworld

Inworld is a product-oriented research lab composed of top AI researchers and engineers. We are dedicated to developing best-in-class real-time multimodal models and operate the only real-time orchestration platform optimized for thousands of queries per second.

We have successfully raised over $125M from prominent investors including Lightspeed, Section 32, Kleiner Perkins, Microsoft’s M12 venture fund, Founders Fund, Meta, and Stanford. Our technology has been instrumental in powering experiences for companies like NVIDIA, Microsoft Xbox, Niantic, Logitech Streamlabs, Wishroll, Little Umbrella, and Bible Chat. We have been recognized by CB Insights as one of the 100 most promising AI companies globally and named one of LinkedIn's Top 10 Startups in the USA.

Who We're Looking For

Reliably working agentic systems are a recent development, and as such, we understand that extensive prior experience in this exact field is rare. We are not seeking candidates who fit a rigid resume template. Instead, we are looking for strong individuals from diverse backgrounds who demonstrate a capacity for rapid learning, thrive in ambiguous environments, and can showcase their past achievements, challenges, and insights.

Experience We Find Useful

While you don't need to possess all of the following, a significant combination of these experiences will strengthen your application:

• Foundation models: Experience with training, new architectures, Reinforcement Learning (RL), reward modeling, and scaling.
• Evaluation: Proficiency in benchmarks, evaluation loops, quality measurement, LLM-as-judge methodologies, and failure analysis.
• Frontier topics: Engagement with multimodal models, agents, tool use, test-time compute, and world models.
• Published research: Contributions to leading conferences such as ICML, ICLR, NeurIPS, EMNLP, ACL, or AAAI.
• PhD in ML/NLP: Or equivalent practical experience that can be demonstrated.
• Public work: Non-trivial AI side projects, interdisciplinary experiments, or open-source contributions.
• Full-stack research ownership: The ability to frame research questions, conduct experiments, author papers, and ship results.

We value learning acquired through building, competitions, or collaborations outside of traditional academia, prioritizing evidence of capability over credentials.

Who Thrives Here

• You are comfortable initiating work without a predefined roadmap, adept at charting the course as you progress.
• You believe that research is incomplete until it has been shipped, demonstrating a bias for impact over purely academic output.
• You delve beyond mere code delivery, obsessing over the why. You are quick to question existing approaches if you identify a superior solution to the core problem.
• You seek deep context and strive to understand the fundamental logic behind every decision, rather than simply accepting instructions.

What Working Here Is Like

We present you with ambiguous problems and expect you to bring clarity to them. We highly value researchers who respond with "I don't know yet" and then design the necessary experiments to find answers. Evaluation is treated as a primary research product, not a mere checklist item before launch. Impact is prioritized over publications, though we fully support sharing work that advances the field. Your contributions will be visible within the company. We operate with a flat organizational structure, emphasize fast iterations, and minimize bureaucratic processes.

We strongly believe in the effectiveness of in-person collaboration for tackling complex problems and fostering a robust team culture. Relocation assistance is available, and we look forward to welcoming you to our Mountain View office.