Lead AI Infrastructure & Accelerator Architect

EPAM Systems; United States; Remote; USD 180,000 - 220,000; Full time

Benefits offered by this job

Medical Insurance

Dental & Vision Insurance

401(k) Matching

Paid Time Off

Job summary

EPAM Systems in the United States seeks a Principal Architect to define the architectural vision for high-performance AI training and serving infrastructure, focusing on TPU and GPU architectures and hardware-software co-design.

You will lead the Center of Excellence, mentor teams, and drive multi-year roadmaps that enable scalable AI across Google Cloud, enterprise customers, and open-source ecosystems. Strong leadership and kernel-level design experience are required.

Qualifications

• BS in Computer Science, Electrical Engineering, or equivalent practical experience (MS/Ph.D. preferred).
• 15+ years of software engineering experience, 8+ years in distributed systems/AI infrastructure/HPC architecture.
• 7+ years designing and developing complex software in C++ or Python.
• 5+ years leading architecture and delivery of large-scale software products or ecosystems.
• Proven kernel-level performance-focused architecture bridging hardware and software.

Responsibilities

• Define multi-year roadmap for high-performance AI kernels and hardware-software co-design for TPU/GPU architectures.
• Scale and mentor a technical practice with governance, standards, and best practices.
• Collaborate with ML researchers, framework architects (JAX, PyTorch), and compiler teams (XLA, MLIR).
• Architect foundational infrastructure incl. benchmarking, autotuning, regression pipelines, and docs.
• Track hardware/ML trends to unlock efficiency in training and inference.

Skills

Distributed systems

AI infrastructure

C++/Python

Architecture leadership

Hardware-software co-design

Education

Bachelor's degree in CS/EE

MS/PhD preferred

Tools

JAX

PyTorch

XLA

MLIR

CUDA

Job description

EPAM Systems in the United States seeks a Principal Architect to define the architectural vision for high-performance AI training and serving infrastructure, focusing on TPU and GPU architectures and hardware-software co-design.

You will lead the Center of Excellence, mentor teams, and drive multi-year roadmaps that enable scalable AI across Google Cloud, enterprise customers, and open-source ecosystems. Strong leadership and kernel-level design experience are required.