Role Description

This role involves accelerating numeric and simulation kernels through GPU/CPU parallelism, memory-hierarchy tuning, and distributed execution across clusters. You’ll design scalable pipelines that maximize efficiency and throughput for large-scale computational workloads.

• Speed up numeric and simulation kernels through GPU/CPU parallelism.
• Optimize workloads via memory-hierarchy tuning and communication reduction.
• Scale computations with MPI, NCCL, and Slurm for distributed clusters.
• Profile and benchmark performance using nvprof and nsys.
• Build reproducible pipelines in Python, NumPy, and SciPy for HPC workflows.
• Collaborate with researchers and engineers to integrate HPC into production AI systems.

Qualifications

• Background in computer science, high-performance computing, or applied mathematics.
• Experienced with GPU/CPU parallel programming using CUDA and OpenMP.
• Understanding of distributed execution frameworks and tools like MPI, NCCL, and Slurm.
• Proficient in Python with libraries like NumPy and SciPy for scientific computing.
• Experience profiling and optimizing workloads with tools like nvprof and nsys.
• Care about memory hierarchy, communication overhead, and scalability in parallel systems.
• Curious about how HPC techniques accelerate AI training, simulations, and scientific workloads.

Requirements

• Design, optimize, and deploy parallel computing pipelines that accelerate numeric, simulations, and large-scale computations across GPUs, CPUs, and distributed clusters.

Benefits

• Classified as an hourly contractor to Mercor.
• Paid weekly via Stripe Connect, based on hours logged.
• Part-time (20–30 hrs/week) with flexible hours—work from anywhere, on your schedule.
• Weekly Bonus of $500–$1000 USD per 5 tasks.
• Remote and flexible working style.