AI Compiler Engineer
The company
After the race to train models, a new challenge has emerged: running artificial intelligence quickly, efficiently, and at scale. To meet this challenge, VSORA, a French semiconductor company founded in 2015, is developing computing accelerators dedicated to AI inference.
Based in Meudon-la-Forêt, VSORA already has a presence in Taiwan, Korea, Japan and the United States, and today represents the only European alternative to NVIDIA and AMD with an architecture aiming for performance more than 3 times greater than existing solutions.
www.vsora.com
Job duties
We are seeking a highly skilled and motivated AI Compiler Engineer to lead the development of the next-generation build chain for the VSORA chip family. This is an operational leadership role: you will join a team of talented engineers, drive architectural decisions, and directly contribute to the implementation of innovative compiler optimizations and merge techniques. You will work at the intersection of AI and systems, helping to shape the future of high-performance AI workloads on VSORA hardware.
Main tasks:
• Participate in the design and development of the AI compilation chain for the VSORA chip family
• Design and implement advanced graph merging and optimization techniques
• Collaborate with hardware architects, ML framework teams, and runtime engineers
• Lead performance analysis and optimization on ML models and benchmarks
• Actively monitor AI compilers, ML frameworks, and hardware acceleration
The desired profile
Profile and skills required:
• Strong C++ and Python programming skills
• In-depth mastery of compiler internals, intermediate representations (IRs), and optimization pipelines
• Hands-on experience with MLIR, TVM, OpenVINO, IREE, XLA
• Good knowledge of PyTorch, TensorFlow, ONNX or other ML frameworks
• Experience in performance optimization and low-level system optimization
• Excellent communication and collaboration skills
Valued features:
• Experience with NPU, GPU, or other AI accelerator architectures
• Contributions to open source projects in the field of compilers or machine learning
• Experience in machine learning or numerical computing
• Proficiency in software development tools (Git, CI/CD, profilers, debuggers)
Apply for job
To view the job application please visit mercato-emploi.com.




NVHBM: Can Custom HBM Change Who Builds the Memory Stack?