Staff AI Systems Engineer - Pre-Training Infra

Reflection; New York (NY); On-site; USD 130,000 - 180,000; Full time

Benefits offered by this job

Top-tier compensation

Comprehensive medical, dental, and vision insurance

Fully paid parental leave

Paid time off

Daily meals provided

Job summary

Reflection, based in New York, is seeking an experienced professional to build and scale distributed training systems for frontier model pre-training. You will work closely with research teams to design large-scale training runs and optimize training efficiency across thousands of GPUs. The ideal candidate has experience with modern distributed training frameworks, strong debugging skills, and familiarity with GPU communication libraries. The company offers top-tier compensation, comprehensive health benefits, and a supportive team environment.

Qualifications

• Experience building or operating distributed training systems for large machine learning models.
• Strong experience with modern distributed training frameworks like Megatron, DeepSpeed.
• Familiarity with large-scale model parallelism strategies.

Responsibilities

• Build and scale distributed training systems for foundation models.
• Design and operate large-scale training runs with research teams.
• Optimize training throughput, stability, and efficiency.

Skills

Distributed training systems

Distributed training frameworks (Megatron, DeepSpeed)

Optimization of GPU utilization

Familiarity with NCCL

Debugging skills in GPU compute

Job description

Reflection, based in New York, is seeking an experienced professional to build and scale distributed training systems for frontier model pre-training. You will work closely with research teams to design large-scale training runs and optimize training efficiency across thousands of GPUs. The ideal candidate has experience with modern distributed training frameworks, strong debugging skills, and familiarity with GPU communication libraries. The company offers top-tier compensation, comprehensive health benefits, and a supportive team environment.