Build Ray-based distributed task graphs for inference, training, RAG, vector search, document retrieval, video, and compound workloads.
Develop Ray head and worker deployment patterns on Kubernetes.
Create per-hardware workload drivers for GPU inference, vector search, storage retrieval, video acceleration, and edge inference.
Implement model-locality and cache-awareness logic so workloads run near the right model or data.
Ensure integration between scheduling decisions, capacity state, SLA rules, billing events, and observability traces.
Conduct end-to-end workload tests for compound RAG and multi-plane AI jobs.

Strong Python distributed-systems experience.
Real Ray experience, preferably production Ray, KubeRay, or distributed AI workload orchestration.
Understanding of AI inference, training, RAG pipelines, model serving, model locality, and accelerator-aware scheduling.
Kubernetes experience for deploying distributed compute systems.
Debugging ability across Python, Kubernetes, GPUs, network movement, and application latency.
Nice to Have: vLLM, TensorRT-LLM, KubeRay, vector search, RDMA storage, GPU scheduling, video pipelines, edge inference, or large-scale model-serving platforms.

Competitive Compensation: Your salary is based on your qualifications, experience, and performance.
Exceptional Stability: Enjoy job security and ample opportunities for career progression.
Professional Development: Access educational programs and certifications to expand your expertise.
Health & Wellness: Comprehensive support for fitness.
Flexible Working Environment: Benefit from a fully remote work setup, flexible schedules, and relocation assistance if needed.
Performance Recognition: Regular bonuses, annual salary reviews, and recognition for your achievements.
Advanced Equipment Options: Choose the workstation setup that fits your needs, whether a desktop or laptop.
Team Building & Community: Participate in regular events that foster collaboration, camaraderie, and innovation.

Distributed AI Scheduling Engineer

Key skills