Latest AI JobsBrowse Jobs
Fireworks AI logo

Member of Technical Staff, Performance Optimization

Fireworks AISan MateoEngineering
Apply on company site
Salary
$175K - $220K USD
Job type
Full-time
Work mode
Remote
Posted
6 May 2025
Status
Open · checked 7d ago

About this role

Fireworks AI is seeking a full-time Member of Technical Staff for Performance Optimization to enhance speed and efficiency across enterprise AI infrastructure. This remote-friendly engineering role is well-suited for professionals with over five years of high-performance computing experience who want to advance generative AI models in San Mateo and beyond.

Role Overview

In this position, the engineer takes complete ownership of performance tuning throughout the entire stack, starting from low-level GPU kernels and moving up to large-scale distributed systems. The professional focuses intensely on accelerating demanding workloads such as large language models, vision-language models, and next-generation video models. Collaboration with research, infrastructure, and systems teams helps eliminate bottlenecks, implement modern optimizations, and scale AI architectures for live production environments.

Key Responsibilities

Enhance system and GPU performance to handle high-throughput AI workloads spanning both training and inference operations.
Investigate and improve latency, memory efficiency, throughput, and compute utilization.
Profile system performance directly to identify and solve GPU and kernel-level bottlenecks.
Develop low-level optimizations utilizing CUDA, Triton, and specialized performance tools.
Boost execution speed and resource management for large-scale architectures like LLMs, VLMs, and video models.
Partner alongside machine learning researchers to co-design and tune model architectures specifically for hardware efficiency.
Expand support for quantization, mixed precision, and model graph optimization.
Establish and maintain robust performance benchmarking and monitoring infrastructure.
Scale training and inference systems across complex multi-GPU and multi-node environments.
Assess and incorporate optimizations designed for emerging hardware accelerators and runtimes.

Requirements

Holds a Bachelor’s degree in Electrical Engineering, Computer Science, Computer Engineering, or possesses equivalent practical experience.
Minimum of 5 years of professional background working directly in high-performance computing systems or performance optimization.
Demonstrated proficiency in CUDA or ROCm alongside practical experience using GPU profiling tools like Nsight, nvprof, or CUPTI.
Familiarity with PyTorch and performance-critical model execution workflows.
Experience debugging and optimizing distributed systems within multi-GPU setups.
Comprehensive understanding of parallel programming models, GPU architecture, and compute kernels.

Nice to Have

Possesses a Master’s or PhD degree in Electrical Engineering, Computer Science, or a related academic discipline.
Prior experience optimizing large-scale models for inference and training, including LLMs, VLMs, or video models.
Working knowledge of machine learning compilers or compiler stacks such as Triton, torch.compile, or XLA.
Past contributions made to open-source high-performance computing or machine learning infrastructure.
Familiarity with cloud-scale orchestration tools and infrastructure like Kubernetes.
Background knowledge emphasizing hardware-aware model design or ML systems engineering.

Skills and Tools

PyTorch
Kubernetes
CUDA / ROCm
Triton
GPU Profiling Tools (Nsight, nvprof, CUPTI)
Distributed Systems

Salary and Benefits

Annual Salary Range: $175,000 to $220,000 USD

About the Company

Fireworks is a specialized intelligence platform designed to help companies build, train, and serve artificial intelligence models customized to their specific data, workflows, and products. Created by the founding team behind PyTorch, the company is backed by prominent investors including AMD, Benchmark Capital, Index Ventures, Lightspeed, NVIDIA, Sequoia Capital, and TCV. Operating as a Series D company valued at $17.5 billion, Fireworks powers production artificial intelligence across text, image, audio, embedding, and multimodal workloads.

How to Apply

Interested candidates should visit the official Fireworks AI careers page to submit an application for this role.

Skills and tech stack

PyTorchKubernetesLLM

Questions candidates ask

What is the salary range for this role?

The annual salary for the Member of Technical Staff, Performance Optimization position ranges from $175,000 to $220,000 USD.

Where is the position located?

The job is based in San Mateo with remote work options available.

What level of experience is required?

Candidates must have 5+ years of experience working on performance optimization or high-performance computing systems, along with a relevant Bachelor's degree or equivalent practical experience.

What specific skills are needed for this position?

Required skills include PyTorch, CUDA or ROCm proficiency, GPU profiling tools experience, distributed system optimization in multi-GPU environments, and a deep understanding of GPU architecture.

What kind of company is Fireworks AI?

Fireworks AI is a Series D enterprise AI platform valued at $17.5 billion that enables companies to build, train, and serve specialized AI models across text, image, embedding, audio, and multimodal workloads.

Ready to apply?

The application is on the official Fireworks AI careers page.

Apply on company site