AI Research Engineer (Multi-Modal & Vision)

Responsibilities

• Conduct end-to-end research and engineering on vision-language models, covering training, evaluation, and optimization.
• Design and implement post-training pipelines including supervised fine-tuning, knowledge distillation, and reinforcement learning from human feedback.
• Develop and maintain high-quality multimodal datasets through curation, filtering, and balancing for domain-specific tasks.
• Drive model efficiency and deployability by adapting models for resource-constrained environments using compression and optimization techniques.
• Design and implement evaluation frameworks and benchmarks to measure model performance, robustness, and real-world task success.
• Build and scale training workflows across distributed GPU infrastructure.
• Stay current with the latest research in multimodal learning and translate findings into practical improvements.
• Publish research findings in top-tier AI conferences and journals.

Requirements

• Degree in Computer Science, Machine Learning, or a related field (MS/PhD preferred).
• Strong experience with multimodal post-training workflows, including supervised fine-tuning and knowledge distillation.
• Hands-on experience with parameter-efficient fine-tuning (PEFT) and distributed training frameworks.
• Demonstrated ability to build and improve vision-language models with measurable results.
• Experience adapting models for resource-constrained environments.
• Proven open-source contributions in multimodal AI on GitHub or HuggingFace.
• Research publications in top AI conferences such as NeurIPS, ICML, ICLR, CVPR, or ECCV.

About the Company

Tether is a global leader in digital finance, pioneering solutions that empower businesses to integrate reserve-backed tokens across blockchains. Through our various divisions—Tether Finance, Power, Data, and Education—we are building the infrastructure for a decentralized future, ranging from the world's most trusted stablecoin, USDT, to cutting-edge AI and peer-to-peer technology.