Hybrid AI Model Developer: Transformer & LLMs

Ericsson GmbH; Austin, Northern (TX, KY); Hybrid; USD 140,000 - 210,000; Full time

Benefits offered by this job

Hybrid work (3 days on-site)

401(k) plan

Health benefits

Job summary

Ericsson in Austin, TX is seeking an AI Model Developer to design, train, and optimize transformer-based models for next-generation telecommunications applications. You will work with researchers, data scientists, and software engineers to deliver production-ready AI capabilities and scalable ML pipelines.

The ideal candidate has 6+ years of industry experience, strong Python and ML framework skills (PyTorch, JAX, TensorFlow), and a deep understanding of model architectures and evaluation.

Qualifications

• 6+ years of industry experience in ML/AI
• Strong experience with PyTorch, JAX, or TensorFlow
• Experience with Transformer models and LLM fundamentals
• Python and modern ML workflows
• Familiarity with ML tooling and version control

Responsibilities

• Design, develop, and optimize transformer-based and deep learning models
• Evaluate architectures and benchmark performance
• Build reference implementations using PyTorch or JAX
• Collaborate with data scientists and software engineers
• Deploy models into production environments
• Mentor junior engineers and uphold engineering practices

Skills

Transformer

Python

ML fundamentals

Communication

Team collaboration

PyTorch

Education

Bachelor3s or Master3s in CS/EE/AI

Tools

PyTorch

JAX

TensorFlow

Hugging Face

MLFlow

ONNX

Ray

Job description

Ericsson in Austin, TX is seeking an AI Model Developer to design, train, and optimize transformer-based models for next-generation telecommunications applications. You will work with researchers, data scientists, and software engineers to deliver production-ready AI capabilities and scalable ML pipelines.

The ideal candidate has 6+ years of industry experience, strong Python and ML framework skills (PyTorch, JAX, TensorFlow), and a deep understanding of model architectures and evaluation.