Remote AI Research Engineer: Agentic Post-Training

Bord; Rosario; A distancia; ARS 185.448.000 - 278.173.000; Jornada completa

Descripción de la vacante

Tether seeks an AI Research Engineer to advance post‑training and agentic workflows for tool‑use models. You will refine pre‑trained models to reason, plan, and autonomously invoke external tools on edge devices, spanning text, images, and audio modalities.

The role emphasizes scalable data pipelines, evaluation, and multi‑node distributed training, with opportunities to publish in top AI venues.

Formación

• Proven expertise in large language model architectures and post-training for agentic workflows.
• Experience with tool use fine tuning, function calling, and RLHF.
• Strong background in data pipelines, evaluation, and multi-turn systems.
• Track record of open-source contributions and publications in AI conferences.

Responsabilidades

• Conduct end‑to‑end research to advance post‑training of agentic and tool‑use models.
• Improve factuality, instruction adherence, tool use, multi‑agent coordination, and reasoning calibration.
• Design scalable post‑training systems, data pipelines, and benchmark infrastructure.
• Develop evaluation suites and diagnostics for model readiness.
• Integrate feedback from real‑world usage into post‑training improvements.
• Collaborate with tooling, product, and training teams to ensure production readiness.
• Coordinate with research and engineering teams to validate integrations for major releases.
• Publish and contribute to openSource repositories related to agentic AI.

Conocimientos

LLM Architectures

Agentic post-training

Tool Use

Function Calling

RLHF

Multimodal

Distributed Training

Research Mindset

Open Source

Publications

Educación

MS/PhD preferred

Herramientas

PyTorch

Hugging Face

GitHub

TensorFlow

Descripción del empleo

Tether seeks an AI Research Engineer to advance post‑training and agentic workflows for tool‑use models. You will refine pre‑trained models to reason, plan, and autonomously invoke external tools on edge devices, spanning text, images, and audio modalities.

The role emphasizes scalable data pipelines, evaluation, and multi‑node distributed training, with opportunities to publish in top AI venues.