CUDA Engineer

This is a fully remote position, open to applicants in Bulgaria.

📋 Description

• Investigate and enhance the performance of LLM inference on contemporary GPUs

• Collaborate with inference and serving frameworks like vLLM, SGLang, TensorRT-LLM, or similar technologies

• Analyze GPU execution, memory access patterns, kernel performance, and identify inference bottlenecks

• Develop, optimize, debug, or reverse-engineer CUDA kernels and low-level GPU operations

• Explore optimization strategies for matrix multiplication, attention mechanisms, and neural network layers

• Benchmark performance and pinpoint opportunities to enhance latency, throughput, and GPU utilization

• Tackle technically challenging issues that necessitate analytical thinking and hands-on experimentation

⛳️ Requirements

• Strong practical experience with CUDA and GPU programming

• Profound understanding of GPU architecture, memory hierarchy, parallel execution, and performance optimization

• Experience with LLM inference, AI systems, GPU kernels, or high-performance computing

• Familiarity with vLLM, SGLang, TensorRT-LLM, Triton, or FlashAttention

• Proven experience in debugging complex performance issues and analyzing low-level system behavior

• Excellent problem-solving abilities

• Capability to convey complex technical concepts in English

• A PhD or research experience in Computer Science, Machine Learning, Computer Architecture, or a related domain (preferred)

• Publications, research projects, or significant contributions to open-source GPU/LLM infrastructure (preferred)

• Experience with PyTorch internals, distributed inference, quantization, KV cache optimization, or custom GPU kernels (preferred)

• Contributions to vLLM, SGLang, Triton, FlashAttention, or related repositories (preferred)

🏝️ Benefits

• Security through thorough client vetting to reduce risks and ensure reliable, timely payments

• Career support and help in finding new opportunities if a project does not align

• Legal assistance regarding independent contractor or sole proprietorship status, taxes, and related processes

• English courses and professional development opportunities

• Team-building activities

• Flexible working hours

• 29 days of paid time off (18 working days per year plus all national holidays)

• 10 paid recovery days

• Comprehensive financial and legal support for independent contractors

• Free English classes with native speakers or Ukrainian instructors

• Dedicated HR support