AI Research Engineer – Agentic Reinforcement Learning

This is a fully remote position, open to applicants in United Kingdom.

📋 Description

• Perform research on reinforcement learning algorithms tailored for multimodal models, focusing on diffusion-based techniques for images, autoregressive models for multimodal comprehension, and integrated multimodal frameworks.

• Architect and develop scalable, distributed reinforcement learning infrastructure for multimodal systems.

• Formulate reward modeling techniques to enhance training stability, align model behaviors, and minimize reward hacking.

• Generate and organize multimodal simulation environments and datasets for training, assessment, and benchmarking purposes.

• Design and implement benchmarking and evaluation protocols across various multimodal tasks.

• Assess and refine policy performance by identifying training bottlenecks, credit assignment issues, and cross-modal alignment challenges.

• Explore next-generation reinforcement learning paradigms that leverage feedback from environments.

• Disseminate research results in prestigious conferences such as ICML, NeurIPS, ICLR, CVPR, ICCV, and ECCV.

⛳️ Requirements

• A Master's degree in Computer Science or a related discipline is mandatory.

• A PhD in Machine Learning, NLP, Computer Vision, or a closely related field is highly desirable.

• Robust history of AI research and publications in leading conferences.

• Proven expertise in executing large-scale reinforcement learning experiments in multimodal and vision-centered systems, including online RL environments.

• Demonstrated influence on domain-specific decision-making and quantifiable advancements in policy performance.

• In-depth knowledge of reinforcement learning algorithms and optimization techniques applicable to vision and multimodal learning challenges.

• Emphasis on policy stability, exploration, and sample efficiency in intricate, high-dimensional environments that incorporate images, videos, and various modalities.

• High proficiency in PyTorch and other deep learning frameworks for vision and multimodal AI.

• Practical experience in constructing end-to-end RL pipelines that encompass simulation, training, evaluation, and deployment within production-grade systems.

• Proven capability to apply empirical research to challenges in multimodal and vision RL, including sample inefficiency, exploration-exploitation dilemmas, and training instability.

• Experience in creating robust evaluation frameworks and iterating on algorithmic enhancements.

• Established record of research publications in leading conferences such as ICML, NeurIPS, ICLR, CVPR, ICCV, ECCV, etc.

• Exceptional English communication skills.

🏝️ Benefits

• Remote work flexibility from any location worldwide.

• Collaboration within a global, distributed team.

• Opportunity to work alongside leading experts in fintech and AI.

• Professional and research prospects through publication in prestigious conferences.

• Engage in innovative projects involving digital finance, blockchain, AI, and peer-to-peer technology.