DevOps / Infrastructure Engineer

3 Month Contract | Remote | High-Impact

About the Role: We’re looking for an infrastructure-focused devOps / MLOps engineer who thrives at the intersection of machine learning, systems, and product delivery. This is a hands-on 3-month contract role responsible for deploying, monitoring, and scaling the testing and deployment infrastructure for a real-time ML-powered content moderation system used to detect and triage abuse, threats, and edge-case language. 

In This Role, You Will:

• Design, build, and document a maintainable GCP cloud infrastructure CI/CD pipeline for real-time model serving and data workflows
• Deploy and optimize APIs for low-latency ML systems
• Automate model deployment, retraining, and evaluation (CI/CD for ML)
• Build observability tooling to monitor rollouts, errors, integration testing, and drift in ML pipelines
• Ensure infrastructure meets security, compliance, and uptime requirements

We’re Looking for Someone Who:

• Has 3–8 years of  DevOps/Platform engineering experience deploying machine learning systems or high-availability backend systems.
• Ability to build CI/CD pipelines from scratch; familiarity with GitHub Actions or similar.
• Expert-level proficiency with Git and GitHub workflows and strong scripting abilities in Python, Bash, and/or Go.
• Experience with Google Cloud Run and Docker. Experience with Google Cloud Platforms, Docker, Kubernetes, Terraform .
• Familiarity with SOC 2 compliance requirements and security best practices (IAM, secrets, etc).
• Experience implementing monitoring, logging, and alerting systems (e.g., Prometheius, Grafana, ELK/EFK, OpenTelemetry).
• Can work cross-functionally with ML, security, and engineering teams to deploy safely and iterate fast.
• Brings a builder's mindset and bias for ownership in ambiguous environments.

What Success Looks Like in the  3 Months:

• You’ve deployed and monitored a real-time ML inference system with well-defined observability.
• You’ve implemented an API with latency under 1000ms for classifier-based inference.
• You’ve partnered with ML engineers to streamline deployment and retraining workflows.
• You’ve built logging and monitoring that gives insight into system performance and classifier behavior.