Relevant Experience & Education Highlights

Expertise in GPU-accelerated deep learning infrastructures, high-performance computing for DNNs, and sparsity optimizations positions this candidate strongly for building scalable foundation model training systems at Adobe.

1. You'll build and optimize infrastructures that power large foundation model training on thousands of GPUs

Extensive background in developing compute support for deep neural networks on GPUs aligns directly with large-scale training requirements.

• Led development of GPU-based distributed mathematical library at a major global electronics company, enabling scalable DNN model applications.
• Optimized scalability and memory usage of distributed computing library while developing batch normalization algorithms for DNNs.
• Advanced research in compute support for deep neural networks on GPUs at a prominent research university laboratory.
• Held senior technical staff roles at a pioneering AI systems company focused on wafer-scale computing for AI workloads.

2. You will profile GPU utilization, trace inference and training runs and help craft strategies for optimizing our ML model latency

Proven track record in analyzing DNN training characteristics and inventing low-overhead monitoring tools supports GPU profiling and latency optimization.

• Investigated data sparsity patterns and trends during DNN training at a leading semiconductor company.
• Invented Sparsity Monitor for DNN training with extremely low data transfer overhead, resulting in a pending patent.
• Led system performance optimizations including power savings and robustness in high-scale communication systems at a top-tier semiconductor corporation.

3. We'll work together to architect and optimize end-to-end ML pipelines, ensuring they're scalable, efficient, and robust

Leadership in designing and resolving issues across development phases demonstrates capability in end-to-end ML pipeline architecture.

• Designed and developed system features for LTE physical layer controller including scheduling, HARQ, and measurements at a leading semiconductor firm.
• Analyzed, identified, and resolved over 300 technical issues during development and verification phases.
• Led development of multiple demo applications using distributed mathematical libraries for DNN models.

Requirements & Candidate Alignment