Senior/Staff AI Graph Compiler Engineer

📋 Description

• • Join Axelera AI, a pioneering deep-tech startup at the forefront of AI innovation, as a Senior/Staff AI Graph Compiler Engineer. We are dedicated to building the next-generation AI platform that empowers advancements in humanity and global problem-solving. In just four years, we've secured $370 million in funding and assembled a world-class team of over 220 employees, including more than 49 PhDs with extensive citation records, operating across 18 countries both remotely and from our offices in Belgium, France, Switzerland, Italy, and the UK, with our headquarters at the High Tech Campus in Eindhoven, Netherlands. Our groundbreaking Metis™ AI Platform is already delivering a 3-5x increase in efficiency and performance, and we have a robust business pipeline exceeding $100 million, solidifying our position as a global industry leader.
• • As a Senior/Staff AI Graph Compiler Engineer, you will be an integral part of our expanding compiler team. Your primary focus will be on the design, development, and optimization of our MLIR-based compiler stack. This critical role enables the efficient execution of complex AI workloads on our state-of-the-art heterogeneous hardware architectures. You will be at the heart of innovation, working closely with AI researchers, fellow compiler engineers, and hardware architects, fostering collaboration within a talented, international team across Europe.
• • This is a unique opportunity to contribute to cutting-edge AI acceleration architectures, push the boundaries of compiler technology, and make a tangible impact within a dynamic and rapidly evolving startup environment. Your contributions will directly influence the performance and capabilities of our AI platform.
• • Key responsibilities will encompass the end-to-end development lifecycle of compiler components. You will be responsible for designing, implementing, and maintaining sophisticated frontend and graph-level compiler modules leveraging the MLIR framework. This includes developing and meticulously optimizing graph-level transformations essential for AI performance, such as advanced operator fusion techniques, intelligent constant folding, strategic operator sinking, efficient graph partitioning, and a suite of other performance-critical optimizations.
• • A significant part of your role will involve extending and maintaining MLIR dialects, passes, and the underlying infrastructure to ensure robust support for diverse and demanding AI workloads. You will be instrumental in integrating and lowering AI models from popular frameworks like PyTorch, ONNX, and TensorFlow into our proprietary internal compiler representations, ensuring seamless compatibility and optimal performance.
• • Close collaboration with our hardware and backend compiler teams will be paramount. You will work hand-in-hand with them to guarantee the efficient and effective mapping of AI workloads onto our heterogeneous hardware architectures, bridging the gap between software and hardware.
• • Beyond development, you will play a crucial role in supporting and mentoring other team members, guiding them in the adoption and effective utilization of our MLIR infrastructure. This includes sharing best practices and fostering a culture of knowledge exchange.
• • You will be tasked with analyzing complex model graphs to identify bottlenecks and implement targeted optimizations that enhance performance, improve memory efficiency, and boost overall execution efficiency. Your insights will be vital in refining our compiler's capabilities.
• • Furthermore, you will contribute significantly to the strategic design and ongoing evolution of our overall compiler architecture and associated tooling, helping to shape the future direction of our technology.
• • Debugging, profiling, and systematically improving the performance and correctness of the compiler will be ongoing responsibilities, ensuring the reliability and efficiency of our AI platform.