Deep Learning Engineer
Core
Build and optimize training infrastructure and frameworks for multi-billion parameter Large Language Models (LLMs), focusing on scaling, performance, and post-training methods.
Role type
Senior IC deep learning engineer (LLM training infrastructure)
Builds
Production-grade LLM training frameworks and serving systems
Domain
Generative AI / Large Language Models / Distributed Systems
Deliverable
production ML models
Required skills
Python, PyTorch, distributed training (FSDP/Megatron/DeepSpeed), LLM architectures (MoE, KV cache), performance optimization, debugging across stack, RL training pipelines, inference optimization
Preferred skills
Kubernetes/Slurm, quantization/distillation, spec-decoding, kernel-level optimization
Technologies
PyTorch, FSDP, Megatron, DeepSpeed, JAX, vLLM, SGLang, TensorRT-LLM, verl, TRL, NCCL
Responsibilities
Develop and maintain training frameworks and infrastructure; optimize training performance and quality via parallelism and architecture tuning; optimize model serving for high throughput and low latency; run training and evaluation experiments; debug distributed training issues and post-training pipelines
Seniority
Senior, hands-on IC
