Job Summary:
Databricks is a leading data and AI company focused on enabling data teams to tackle complex problems. As a Senior GenAI Research Engineer, you will drive performance improvements and design high-performance GPU kernels for large language model training while collaborating with a diverse team of researchers and engineers.
Responsibilities:
โข Drive performance improvements through advanced optimization techniques including kernel fusion, mixed precision, memory layout optimization, tiling strategies, and tensorization for training-specific patterns
โข Design, implement, and optimize high-performance GPU kernels for training workloads (e.g., attention mechanisms, custom layers, gradient computation, activation functions) targeting NVIDIA architectures
โข Design and implement distributed training frameworks for large language models, including parallelism strategies (data, tensor, pipeline, ZeRO-based) and optimized communication patterns for gradient synchronization and collective operations
โข Profile, debug, and optimize end-to-end training workflows to identify and resolve performance bottlenecks, applying memory optimization techniques like activation checkpointing, gradient sharding, and mixed precision training.
Qualifications:
Required:
โข BS/MS/PhD in Computer Science or related field with hands-on experience writing and tuning CUDA kernels for ML training applications, or hands-on experience in distributed training frameworks (PyTorch DDP, DeepSpeed, Megatron-LM, FSDP)
โข Strong understanding of NVIDIA GPU architecture (memory hierarchy, tensor cores, warp scheduling, SM occupancy) and proficiency with CUDA debugging/profiling tools (Nsight, NVProf)
โข Deep understanding of parallelism techniques and memory optimization strategies for large-scale model training, with proven ability to debug and optimize distributed workloads
โข Strong software engineering skills in Python and PyTorch, with experience supporting production training workflows and knowledge of LLM training dynamics including hyperparameter tuning and optimization strategies.
Company:
Databricks is a data and AI platform that unifies data engineering, analytics, and machine learning on a lakehouse architecture. Founded in 2013, the company is headquartered in San Francisco, USA, with a team of 5001-10000 employees. The company is currently Late Stage.