AI Platform Engineer, Training and Inference

Аналитика рынка

Стек

Откликнуться

Загрузите резюме — мы свяжем вас с работодателем напрямую через нашу базу.

Описание

AI Platform Engineer – Training & Inference Saviynt's AI-powered identity platform manages and governs human and non-human access to all of an organization's applications, data, and business processes. Customers trust Saviynt to safeguard their digital assets, drive operational efficiency, and reduce compliance costs. Built for the AI age, Saviynt is today helping organizations safely accelerate their deployment and usage of AI. Saviynt is recognized as the leader in identity security, with solutions that protect and empower the world's leading brands, Fortune 500 companies and government institutions. For more information, please visit [link] The AI Platform team is building the compute layer that trains, evaluates, and serves every AI model at Saviynt. We need an ML Platform Engineer to own distributed training on Ray + H100s, the multi-engine LLM inference mesh (vLLM, SGLang, NVIDIA Triton), and the full model promotion lifecycle — from shadow mode through canary rollout to GA. The AI Platform team's mission is to build a secure, scalable, product-agnostic AI foundation that enables Saviynt's identity products to deliver measurable AI-powered outcomes. Training & Inference is the engine — it turns data into deployed models that make Saviynt's products smarter.   What You Will Be Doing • Own the Ray ecosystem end-to-end: manage KubeRay on GKE, tune Ray Core Task/Actor scheduling, operate the Plasma distributed object store, and configure Ray Data for GPU-direct streaming from GCS/S3 • Operate distributed training with Ray Train: configure TorchTrainer + DDP/NCCL for multi-node H100 clusters, manage checkpoint lifecycle, implement spot-preemption recovery, and integrate warm-start fine-tuning for retrain pipelines • Build and operate the LLM inference mesh with Ray Serve: compose vLLM (PagedAttention), SGLang (RadixAttention), and NVIDIA Triton (TensorRT/ONNX) as a unified deployment graph with Plasma zero-copy memory sharing • Optimise inference perfor

Контакты работодателя (email/phone/telegram) скрыты из публичного превью — отправьте резюме, чтобы мы связали вас напрямую.