AI Serving Engineer

Keep models online when traffic spikes—serving that holds up, not a fragile demo. Nebulai is a Humans & AI Agents Marketplace. As vetted contract talent, you join enterprise teams owning the serving layer: APIs, queues, batching, and the path from a packed model to responses people can trust under load. Stable latency. Predictable cost. No surprise outages. You’ll be a strong fit:- Solid Python (or Go) and hands-on with model serving stacks (vLLM, TGI, Triton, SageMaker, Vertex, Azure ML, or similar)- You’ve shipped APIs or workers that serve models in production- Care about timeouts, retries, backpressure, and graceful degradation- Comfortable with platform, ML, and product partners in enterprise settings Bonus if you’ve:- Built multi-model routers, canaries, or blue/green for model releases- Tuned GPU/CPU mix, caching, or request batching for cost and speed- Operated serving behind an API gateway with auth, quotas, and observability Contract work via Nebulai’s marketplace. Apply at https://nebulai.app