Software Engineer - Training Product
Core
Build multi-node training and serverless reinforcement learning features for AI companies to accelerate model development and deployment.
Role type
Senior IC software engineer (ML training infrastructure)
Builds
Multi-node training pipelines, serverless RL products, checkpointing systems, and developer tooling for model training.
Domain
AI/ML infrastructure, distributed systems, cloud computing
Deliverable
production ML models | product features
Required skills
distributed systems, web stack, database design, API design, infrastructure development, model fine-tuning, Kubernetes, product sense for developer tools
Preferred skills
feature lifecycle management (MVP to GA), model development paradigms (SFT, RL, LoRA), open-source training frameworks (PyTorch, Megatron, DeepSpeed), frontend development
Technologies
Kubernetes, vLLM, NCCL, PyTorch, Megatron, NemoRL, VeRL, Axolotl, HF Trainer, FSDP, DeepSpeed
Responsibilities
Architect solutions from API/UI to infrastructure layer, fine-tune and deploy models to understand workflows, partner with research engineers on post-training requirements, drive system reliability and development velocity, fix bugs and resolve customer issues
Seniority
Senior, hands-on IC