AI Research Engineer (Agentic Reinforcement Learning)
New
T
TetherAI research
Source API remote eligibility restrictions: Argentina, Bangladesh, Belgium, Brazil, Bulgaria, Colombia, Cyprus, Czechia, Denmark, Estonia, Georgia, Greece, Hungary, India, Ireland, Israel, Japan, Malta, Netherlands, Norway, Pakistan, Poland, Portugal, Romania, Serbia, Spain, Sweden, Switzerland, Taiwan, Thailand, Turkey, United Arab Emirates, United Kingdom, Uruguay, VietnamFull-Time
Salary not disclosed
Apply NowOpens the employer's application page
Job Details
- Languages
- Excellent English communication skills
- Required Skills
- PyTorchDeep Learning
Requirements
- Hold a Master’s degree in Computer Science or a related field.
- A PhD in Machine Learning, NLP, Computer Vision, or a closely related discipline is preferred.
- Have a strong track record of AI research and publications in top-tier conferences.
- Bring experience running large-scale reinforcement learning experiments in multimodal and vision-centric systems, including online RL settings.
- Demonstrate impact on domain-specific decision-making and measurable improvements in policy performance.
- Understand reinforcement learning algorithms and optimization methods for vision and multimodal learning, including policy stability, exploration, and sample efficiency.
- Be proficient in PyTorch and deep learning frameworks for vision and multimodal AI.
- Have hands-on experience building end-to-end RL pipelines covering simulation, training, evaluation, and production deployment.
- Have experience addressing sample inefficiency, exploration-exploitation tradeoffs, and training instability in multimodal and vision tasks.
- Have experience designing robust evaluation frameworks and iterating on algorithmic improvements.
Responsibilities
- Conduct research on reinforcement learning algorithms for multimodal models, including diffusion-based approaches, autoregressive models, and unified multimodal frameworks.
- Design and build efficient, reliable RL infrastructure for scalable distributed training across multimodal systems.
- Develop reward modeling strategies to improve training stability, align model behavior with desired outcomes, and mitigate reward hacking.
- Create and curate multimodal simulation environments and datasets for training, evaluation, and benchmarking.
- Design benchmarking and evaluation protocols to measure model performance, track baselines, and validate improvements across multimodal tasks.
- Analyze and optimize policy performance across modalities by identifying bottlenecks in training, credit assignment, and cross-modal alignment.
- Investigate reinforcement learning paradigms that learn more effectively from environment feedback.
- Publish research findings in top-tier conferences such as ICML, NeurIPS, ICLR, CVPR, ICCV, and ECCV.
View Full Description & ApplyYou'll be redirected to the employer's site
