Tasks:
- * Degrade to fully on-device inference on poor networks
- * Design split-computing inference using data compression or feature extraction
- * Develop semantic-complexity and privacy-based routing algorithms
- * Implement on-device PII filtering before cloud requests
- * Monitor network conditions and optimize workload allocation
- * Preserve context when tasks transition between local and cloud execution
- * Route simple tasks to on-device models and complex reasoning tasks to cloud models
- * Synchronize agent memory and task execution state across device and cloud
- * Train model routers and classifiers
Perks/Benefits:
Skills/Tech stack required:
[5G] [Agent memory] [Agent Memory Management] [Classifier Training] [Claude 3.5] [Claude 3.5 API] [Cloud Inference] [Consistency protocols] [Data Compression] [Device Inference] [Distributed Systems] [Feature extraction] [Gemini API] [GPT-4o] [GPT-4o API] [Language Models] [Large Language Models] [Memory Management] [Model routing] [Network-aware workload optimization] [Network handling] [On-device Inference] [PII filtering] [Privacy-preserving] [Privacy-Preserving AI] [Semantic routing] [Small Language Models] [Small Language Models (SLMs)] [Split Computing] [State synchronization] [Weak Network] [Weak-network handling] [Wi-Fi] [Workload optimization]
Educational requirements:
N/A
Role(s):
[AI Architect] [AI Systems Architect] [Architect] [Hybrid AI Architect] [Systems Architect]