12 hours ago
London, United KingdomSenior
Responsibilities
- Own and deliver projects on the team roadmap that support high-impact technical and business milestones.
- Design and implement the compiler, runtime, and debugger for PyTorch, TensorFlow, JAX, and MXNet on custom hardware.
- Optimize performance using layout, fusion, scheduling, and tiling techniques to maximize throughput.
- Lead code reviews, coach peers, and uphold engineering standards for ML backend and performance engineering.
- Track GPU and AI accelerator trends and propose and prototype innovations.
- Collaborate with hardware and modeling teams on device trade-offs and model performance.
Requirements
- 5+ years of software engineering experience focused on C/C++ programming.
- Deep expertise in ML framework internals, compilers, low-level programming, and optimization techniques.
- Deep expertise optimizing TensorFlow, PyTorch, or JAX deep-learning models.
- Deep expertise with multiple toolchains such as LLVM, OpenXLA/XLA, MLIR, and TVM.
- Practical experience applying machine learning in high-performance computing contexts.
- Bachelor’s degree in computer science, electrical engineering, telecommunications engineering, mathematics, or a related field.
- Strong problem-solving, critical-thinking, teamwork, communication, and cross-functional collaboration skills.
- Experience working in fast-paced, dynamic environments.
- Personal projects are considered an important differentiating factor.
Benefits
- Meaningful stock options and equity ownership.
- Employer-contributed retirement plans.
- Annual Living-Local Bonus for employees living within 20 minutes of the office.
- Competitive salary commensurate with experience, skills, and location.
- Eligibility to work at OLIX is subject to U.S. export-control restrictions based on most recent citizenship or permanent-residency status.
About OLIX
OLIX designs specialized semiconductors and rack-scale systems to accelerate AI token production, re-architecting data centers with dedicated machines, fast interconnects, and optics. Its first product, the Decode Accelerator 1 (DX-1), targets the decode stage of large-model inference to improve throughput and efficiency for hyperscalers and AI platform operators. Privately held and founded in 2024, OLIX recruits across London, Bristol, Austin, Toronto, and San Francisco.