Principal AI Accelerator Software Architect

Ampere Computing LLC; Warszawa; Hybrid; PLN 302,000 - 453,000; Full time

Benefits offered by this job

Premium medical health care

Generous paid time off policy

Office amenities and free gym access

Remote work policy with connectivity/e

Sports card

Job summary

Ampere Computing LLC in Poland seeks an AI Accelerator Software Principal Engineer – Framework Integration to lead end-to-end strategy and delivery for high-performance deep learning inference across Ampere accelerator platforms. You will shape how PyTorch, ONNX and llama.cpp are enabled and optimized for our hardware, delivering high throughput and low latency for data centers to edge.

The role requires deep ownership, cross-team influence, and mentoring of engineers while advancing runtime,

Qualifications

• Must have strong background in deep learning frameworks and integration.
• Experience optimizing AI workloads on accelerator backends.
• Proficiency in Python and C/C++ with performance-focused engineering.

Responsibilities

• Lead integration of PyTorch, ONNX, llama.cpp into Ampere backend.
• Drive end-to-end acceleration across SW/HW stack including inference serving, runtime, and kernels.
• Mentor engineers and establish engineering standards for performance and maintainability.

Skills

Deep learning

Framework integration

Performance engineering

Python programming

C/C++ programming

Education

BS/MS/PhD in CS/EE/CE or related

Tools

PyTorch

ONNX

llama.cpp

Linux drivers/runtime

Job description

Ampere Computing LLC in Poland seeks an AI Accelerator Software Principal Engineer – Framework Integration to lead end-to-end strategy and delivery for high-performance deep learning inference across Ampere accelerator platforms. You will shape how PyTorch, ONNX and llama.cpp are enabled and optimized for our hardware, delivering high throughput and low latency for data centers to edge.

The role requires deep ownership, cross-team influence, and mentoring of engineers while advancing runtime,