Ai Engineer • Mid-Level

Who you are

You have deep systems or infrastructure engineering experience in high-performance or distributed computing environments — you've designed and built reliable systems that operate at scale, ensuring optimal performance and efficiency. Your strong understanding of GPUs, CUDA, and NCCL allows you to push the limits of hardware and software to achieve remarkable results in large-scale training and inference.

Hands-on experience with containerization and orchestration tools like Kubernetes or Slurm is essential — you've managed clusters and optimized resource utilization, ensuring that thousands of GPUs are effectively utilized. You thrive in a fast-paced, high-ownership startup environment, demonstrating high agency and the ability to collaborate closely with cross-functional teams.

Your familiarity with modern observability stacks and performance profiling tools enables you to monitor and enhance system performance — you understand the importance of observability in maintaining reliability and efficiency in complex systems. You are passionate about building tools and automation for distributed training and inference, contributing to the acceleration of AI research and development.

Desirable

Experience with large-scale training frameworks and libraries such as PyTorch, DeepSpeed, JAX, and Megatron-LM is a plus — you are eager to learn and adapt to new technologies that can enhance system capabilities. A background in AI research or collaboration with research teams will further enrich your contributions to the team.

What you'll do

In this role, you will design, build, and operate Reflection’s large-scale GPU infrastructure — your work will power pre-training, post-training, and inference for cutting-edge AI models. You will develop reliable, high-performance systems for scheduling, orchestration, and observability across thousands of GPUs, ensuring that the infrastructure meets the demands of AI research and deployment.

Your responsibilities will include optimizing cluster utilization, throughput, and cost efficiency while maintaining reliability at scale — you will push the limits of hardware, networking, and software to accelerate the path from idea to model. Collaborating closely with research, training, and platform teams, you will enable large-scale training and inference, contributing to the advancement of open superintelligence.

You will build tools and automation for distributed training, inference, monitoring, and experiment management — your contributions will directly impact the efficiency and effectiveness of AI research at Reflection. You will also engage in performance profiling and optimization, ensuring that the infrastructure remains robust and responsive to the needs of the team.

What we offer

At Reflection, we provide a supportive and inclusive work environment where you can thrive — we offer fully paid parental leave for all new parents, including adoptive and surrogate journeys, and financial support for family planning. Our benefits include paid time off when you need it, relocation support, and perks that optimize your time.

You will have opportunities to connect with teammates through daily lunches and dinners, as well as regular off-sites and team celebrations — we believe in fostering a strong team culture that encourages collaboration and innovation. Join us in our mission to build open superintelligence and make it accessible to all.