Lucas Gonthier

Moving from a promising notebook to a reliable, scalable, and observable ML system in production is rarely about the model. It's an engineering challenge. That's my job.

Machine Learning & MLOps Engineer, I support clients in three areas:

Design of custom ML architectures, aligned with business constraints, data volumes, and performance requirements.

Design, deployment, and operation of scalable and reliable ML systems.

I view models as components of a distributed system, not isolated artifacts: observability, drift management, lifecycle automation, infrastructure as code, ML CI/CD.

Everything that makes a system hold up under real load, without manual intervention at 3 AM.

Standardization of deployment processes, upskilling ML teams, drastic reduction of time-to-market.

-ML / DL: PyTorch, JAX, XGBoost, Optuna

-MLOps: FastAPI, Kubeflow, MLflow, KServe, Evidently

-Infra & Cloud: Kubernetes, GCP, Terraform, Helm, GitHub Actions

-Agents / LLMs: LangGraph, LangSmith

Stack adaptable to project context. I align with the team's existing setup.

- You have an ML prototype that needs to go into production without breaking at the first load increase

- Your ML team lacks MLOps structure and wastes time on deployment

- You want to build a scalable ML or agentic system from scratch

- You're looking for someone who codes, architects,andtransfers knowledge to the team

Let's discuss.