Remote AI Systems Architect: Edge Inference & Performance

Anyone Ai Inc.; Buenos Aires; Remote; ARS 184,482,000 - 353,590,000; Full time

Job summary

Anyone Ai Inc. is seeking an AI Research Engineer to advance kernel and inference optimization for scalable AI systems. You will design and optimize model serving pipelines, work on low-latency inference for edge devices, and collaborate with cross-functional teams to productionize advanced serving techniques.

The role emphasizes GPU kernel work on mobile devices, MSL shaders, and empirical research to push inference performance while managing memory footprints in constrained environments.

Qualifications

• Design and optimize state-of-the-art model serving architectures for high throughput and low latency.
• Develop robust inference pipelines with measurable performance metrics.
• Experience with mobile device optimization and edge deployments.
• PhD in NLP/ML and publications are a plus.

Responsibilities

• Design and deploy model serving architectures for high throughput and low latency.
• Build and monitor inference tests in simulated and live environments.
• Identify test datasets and scenarios for low-resource devices.
• Collaborate with cross-functional teams to productionize optimized frameworks.
• Analyze bottlenecks and optimize memory usage on edge platforms.

Skills

Model serving

Inference optimization

GPU kernels

Edge devices

Low-level optimization

Distributed inference

Education

PhD in NLP/ML or related field

Computer Science degree or related field

Tools

MSL

Custom compute shaders

Model serving frameworks

Job description

Anyone Ai Inc. is seeking an AI Research Engineer to advance kernel and inference optimization for scalable AI systems. You will design and optimize model serving pipelines, work on low-latency inference for edge devices, and collaborate with cross-functional teams to productionize advanced serving techniques.

The role emphasizes GPU kernel work on mobile devices, MSL shaders, and empirical research to push inference performance while managing memory footprints in constrained environments.