About the role
The gap between a model that works in a notebook and one that runs reliably in a client’s stack is where most AI projects stall. As an MLOps Engineer at ASC, you close that gap: you build the deployment, monitoring, and retraining infrastructure that lets a model live in production and stay accurate over time.
You will set up the pipelines, the observability, and the handover artifacts on every engagement, and you will make sure the client can operate what we ship without us in the room. This role sits between the ML engineers and the client’s platform team.
What you'll do
- Build CI/CD and deployment pipelines for models and the services around them.
- Stand up monitoring for model performance, drift, latency, and cost, with alerts that mean something.
- Automate retraining and evaluation so models stay accurate as the client’s data changes.
- Containerize and orchestrate workloads (Docker, Kubernetes) across cloud platforms.
- Package the infrastructure, runbooks, and dashboards so the client can run the system after handover.
What we're looking for
- 4+ years in DevOps, platform, or ML infrastructure, with production ownership.
- Strong Docker and Kubernetes, plus infrastructure-as-code (Terraform or equivalent).
- Experience with an ML platform or tooling (MLflow, Kubeflow, SageMaker, Vertex AI, or similar).
- Fluency with at least one major cloud (AWS, Azure, or Google Cloud) and its cost model.
- A reliability mindset: you design for the 3am page, not the demo.
Nice to have
- Experience serving LLMs or large models efficiently (batching, quantization, GPU scheduling).
- Security and compliance automation in regulated environments.
- Working proficiency in Korean or Japanese.
How we hire
- 1
Intro call
A short conversation about your background, what you want to work on, and how ASC works.
- 2
Technical conversation
A working session on real problems in your area, closer to how we actually work than to a quiz.
- 3
Team and offer
Meet the people you would work with, then a decision, usually within a week.