AI Research Engineer, Multimodal Vision at Tether

Role overview

This AI model role covers the full vision-language development cycle, from curated datasets and training pipelines to evaluation and deployment efficiency. The employer seeks measurable improvements for real-world and resource-limited use. Each requisition states its own remote location.

Responsibilities

• Research, train, evaluate and optimise vision-language models.
• Build post-training pipelines using supervised fine-tuning, knowledge distillation and reinforcement learning from feedback.
• Curate, filter and balance multimodal datasets for domain-specific tasks.
• Compress and optimise models for constrained hardware; define benchmarks for quality, robustness and real-world performance.
• Scale training on distributed GPUs, resolve pipeline bottlenecks and contribute relevant open-source research and tools.

Qualifications

• Computer science, machine learning or related degree; a master’s or PhD is preferred.
• Experience with multimodal post-training, parameter-efficient fine-tuning and distributed training.
• Demonstrated improvements to vision-language models on benchmarks or real-world tasks.
• Experience adapting models to constrained environments, plus open-source contributions and AI research publications as listed by the employer.

About Tether

Tether develops digital-token products, including USDT, alongside projects in peer-to-peer communications and AI. The employer describes a globally distributed remote team.

Related jobs

• AI Enablement Engineer — Keyloop
• AI Product Builder — Money Forward Vietnam
• Principal Engineer — Keyloop
• QC Engineer — VSOL
• Senior Adaptive AUTOSAR Engineer — VinAI
• Senior Boomi Developer — Keyloop