Software Engineer - Compiler

As a Software Engineer - Compiler at CentML, your key responsibilities include writing modular, high-quality code in Python and C++, implementing compiler optimizations, developing efficient GPU kernels with Hidet, and collaborating with peer teams to enhance AI workload performance. Your job will also involve creating tests and benchmarks to validate the solutions we implement.

To qualify for the Software Engineer - Compiler position at CentML, candidates should hold a Bachelor’s or higher degree in Computer Science or a related field. Strong coding proficiency in Python, C++, CUDA, or OpenCL is crucial, along with at least two years of experience in compiler development or performance engineering, preferably in relation to AI workloads.

A Software Engineer - Compiler can stand out at CentML by having experience with ML systems such as TensorFlow or PyTorch, familiarity with ML compilers like TVM and LLVM, and a strong publication record in related fields. Contributions to open-source projects will also enhance your profile significantly.

At CentML, the work environment is open and inclusive, fostering collaboration and innovation. The company promotes a healthy work-life blend with flexible vacation time and professional development budgets. Each team member is encouraged to contribute actively to a shared mission of democratizing AI.

CentML offers a wide array of benefits for Software Engineer - Compiler positions, including employee stock options, comprehensive medical and dental benefits, a parental leave top-up for six months, and a professional development budget. The company also ensures equal opportunities for everyone, embracing diversity and making reasonable accommodations for applicants with disabilities.

When answering this question, emphasize a specific project where you improved performance through optimization techniques. Discuss the tools and methods you used, the challenges faced, and how your efforts led to measurable improvements.

Share your systematic approach to debugging, such as isolating the problem, using tools for diagnostics, and testing various parts of the code. Illustrate your methodology with examples to provide depth to your response.

Discuss your regular engagement with industry literature, workshops, conferences, and online forums. Mention any specific communities or resources that you find particularly beneficial in staying informed about the latest trends.

Highlight any projects where you've used CUDA or OpenCL. Discuss the specifics of how you exploited these technologies to achieve high performance and address any challenges you encountered.

Approach this question by outlining the project’s challenges, your specific contributions, and the successful outcomes. Focusing on your problem-solving skills and teamwork will showcase your capabilities effectively.

Detail your process for writing tests, including defining success criteria, ensuring comprehensive coverage, and methodically analyzing performance results. Present any examples from past experiences where your testing contributed to project success.

Showcase your knowledge of the Hidet compiler, its positioning within the AI development landscape, and your enthusiasm for contributing to its improvement. Share specific features or capabilities you find compelling.

Discuss your approach to prioritization, such as assessing urgency, impact on project timelines, and stakeholder input. Highlight any tools or techniques you use to manage your workload effectively.

Illustrate your perspective on teamwork by highlighting experiences where collaboration led to successful outcomes. Emphasize communication, knowledge sharing, and respect for diverse ideas in the engineering process.

Describe your learning process, including resources you use, how you apply new knowledge, and examples of technologies you’ve adapted to quickly. Your response should demonstrate a proactive approach to continuous learning.