About the role
Where multiple locations are listed for this role, the position may be based in any of those locations, with priority determined according to the order of listing.
What you’ll do
As a Research Scientist at Simular, you will:
- Shape the future of agentic AI by pioneering new research directions in planning, reinforcement learning, multimodal reasoning, grounding, human-agent. interaction, and alignment (e.g. reward modeling, automated task evaluation, AI safety).
- Design and execute experiments end-to-end: from data collection and benchmarking, to model training and evaluation.
- Develop new methodologies that push the boundaries of AI agent capabilities and reliabilities, advancing both academic and product impacts, in collaboration with world-class scientists and engineers.
- Collaborate closely with engineers to integrate research prototypes into product
- Publish and present results at top-tier conferences, contributing to the broader AI research community.
- Advance the science hands-on through designing methods, building datasets, running experiments, and benchmarking against state of the art.
You might be a fit if
- PhD in Computer Science, Machine Learning, or related field, or equivalent research experience.
- Experience owning the complete research lifestyle: leading and executing a research program from ideation all the way to publication.
- Expertise in one or more of (but not limited to) the following areas:
- Pre-training, post-training, or fine-tuning LLMs/VLMs
- Reinforcement learning and/or LLM-based agents
- Computer vision and multimodal learning
- Representation learning and embeddings.
- Consistent publication record as first author in leading conferences (NeurIPS, ICML, CVPR, ACL, etc.).
- Strong implementation skills in modern ML frameworks (PyTorch, JAX, etc.).
- Demonstrated ability to run complex experiments, analyze results, and iterate quickly.
- Passion for high-impact research with real-world applications.
- Ambition and vision to redefine what’s possible in AI research and deployment.