Compiler Engineer | AI Hardware Solutions

Role Overview

Charterhouse is partnering with a rapidly growing technology innovator at the forefront of AI hardware. Our client designs customized server infrastructure and NPUs optimized specifically for machine learning applications. As a Compiler Engineer, you will be the vital link between high-level AI frameworks and raw silicon performance. You will design and implement the software stack that translates complex neural networks into high-performance, energy-efficient machine code, directly influencing the next generation of global AI infrastructure.

Key Responsibilities

1. MLIR Dialect & Pipeline Design

Dialect Development: Design and implement custom MLIR dialects and lowering pipelines tailored for advanced AI accelerator and NPU platforms.
Graph Transformation: Transform machine learning graphs from frameworks like ONNX and PyTorch into optimized kernels.
Optimization Strategies: Apply advanced compiler techniques, including operator fusion, tiling, and vectorization, to maximize hardware utilization.

2. Performance Engineering & Benchmarking

Profiling & Tuning: Profile and benchmark compiler output to ensure it meets strict targets for latency, throughput, and power efficiency.
Hardware-Software Co-design: Collaborate closely with hardware architects to influence NPU features and with runtime teams to co-design compiler-driven execution models.
Portable Architecture: Build a modular and portable compiler stack that provides consistent support across multiple generations of AI hardware.

3. ML Framework Integration

Front-end Mapping: Develop lowering paths for TensorFlow, PyTorch, and ONNX, ensuring seamless translation of high-level operations into efficient IR.
Core Operations: Optimize fundamental ML operations, including tensor manipulations, matrix multiplications (GEMM), and quantization techniques.
Dynamic Handling: Implement robust solutions for handling dynamic shapes and variable workloads within the compiler stack.

About You: Skills & Competencies

C++ Mastery: Expert-level proficiency in modern C++ (14/17/20) for high-performance systems development.
Compiler Internals: A deep understanding of Intermediate Representations (IR), instruction scheduling, and sophisticated loop transformations.
Problem Solver: Confident in tackling the unique challenges of NPU memory movement and resource sharing.
Collaborative Engineer: Ability to work across the stack, from hardware architects to framework developers, to deliver an integrated solution.

Qualifications & Experience

Education: Degree in Computer Science or Computer Engineering; an advanced degree (M.S. or Ph.D.) specializing in Compilers or Parallel Computing is highly preferred.
Professional Experience: A minimum of 5 years in compiler engineering or high-performance computing.
Technical Stack: * Hands-on experience with MLIR or LLVM (specifically dialect design and lowering).
- Proficiency in Python for scripting, testing, and tooling.
- Familiarity with NPU/TPU architectures and AI accelerator hardware.
Differentiators: Contributions to open-source projects (e.g., LLVM, Torch-MLIR, ONNX-MLIR) and experience with memory movement engines and runtime scheduling.

Why Join This Innovation Leader?

This role offers the chance to work in a high-stakes, “greenfield” environment where your code directly dictates the efficiency of AI at scale.

Cutting-Edge Hardware: Gain early access to proprietary NPU designs and influence their evolution.
Global Impact: Contribute to an energy-efficient AI mission that accelerates research and deployment worldwide.
Dynamic Environment: Join a lean, agile team that values technical excellence and rapid iteration over legacy bureaucracy.

Are you ready to bridge the gap between AI frameworks and custom silicon? Apply today to join this technology innovator as a Compiler Engineer.

More information

Job Location : Dubai, United Arab Emirates
City : Dubai
Company Name : Charterhouse

More Jobs

To apply for this job please visit www.charterhouseme.ae.