Principal Model Researcher

Ericsson

Austin, United States
On-site
Transformer architecture mastery
Hardware-aware ml thinking
Jax/pytorch export/compilation
Shape hardware around the model before a single transistor is placed, occupying the most strategically critical seat in the silicon program

Job Summary

  • Shape hardware around the model before a single transistor is placed, occupying the most strategically critical seat in the silicon program.
  • Take bleeding-edge AI and RAN algorithms and convert them into precise hardware specifications and concrete lowering requirements.
  • Project performance using cycle-accurate simulators and SystemC models to give the silicon team confidence for multi-million dollar tape-out decisions.

Matching Summary

Shape hardware around the model before a single transistor is placed, occupying the most strategically critical seat in the silicon program.

Skills & Requirements

Must-have

  • Transformer architecture mastery
  • Hardware-aware ML thinking
  • JAX/PyTorch export/compilation
  • Cycle-accurate performance projection
  • Model partitioning and tiling strategies

Nice-to-have

  • Applied AI to RAN workloads
  • Understanding of MLIR compiler journey
  • Experience with complex-valued AI models

Key Requirements

  • Deep knowledge of Transformer architectures
  • Experience with hardware-in-the-loop
  • Advanced proficiency in JAX, PyTorch, or TensorFlow
  • Performance modeling experience (SystemC, TLM)
  • Experience with real RAN workloads (preferred)
  • Understanding of MLIR transformation passes (preferred)
  • Hands-on experience with complex-valued AI models (preferred)

Work Rights

Not specified

Tailored Resume

Cover Letter