Staff MLOps Engineer / ML Platform Architect

Our client is a technology-driven organization based in Singapore, scaling its AI capabilities across core products. They are investing in a robust ML Platform to move from experimentation to reliable, cost-efficient production.

The Opportunity

Our client is looking for a Senior MLOps Engineer to be the key engineer responsible for productionizing ML and GenAI workloads. This is a senior Individual Contributor role - high ownership, deep hands-on, working directly with Data Science, Data Engineering, and Platform teams.

You will be the go-to expert for how we ship models to production reliably.

Key Responsibilities

1. ML Lifecycle & Productionization

• Build and maintain CI/CD pipelines for ML: training, validation, versioning, deployment, and automated rollback
• Productionize batch and real-time models, including LLM / RAG based services
• Manage Model Registry, Experiment Tracking (MLflow / W&B), and Feature Store

2. ML Platform & Infrastructure

• Build scalable ML infrastructure on AWS / GCP using Docker, Kubernetes (EKS / GKE), Terraform
• Implement Infrastructure as Code (Terraform) and GitOps best practices
• Optimize compute - GPU/TPU, spot instances, inference cost and latency

3. Reliability, Observability & Governance

• Implement monitoring for data drift, concept drift, model performance, and pipeline health using Prometheus / Grafana / CloudWatch
• Ensure reproducibility, data/model versioning, security, and auditability
• Partner with Security & Compliance on model governance and Responsible AI practices (PDPA)

4. Collaboration

• Work closely with Data Scientists to optimize training jobs and inference pipelines
• Work with Data Engineers on feature pipelines and data validation
• Define and document MLOps best practices, templates, and runbooks

Requirements

Must-Have:

• Experience in Software Engineering / DevOps / MLOps, with at least 2+ years focused on MLOps
• Strong Python, solid software engineering (Git, CI/CD, unit testing, code reviews)
• Hands-on experience taking ML models to production in a commercial setting
• Strong Cloud experience (AWS preferred: SageMaker, S3, EKS, Lambda) + Kubernetes + Docker
• Experience with orchestration: Airflow / Prefect / Kubeflow
• Experience with MLflow, model packaging, and model serving (REST/gRPC, Seldon / BentoML / FastAPI)

Good to Have:

• Experience with Feature Store (Feast) and Vector DBs (Pinecone, Weaviate, pgvector) for RAG
• LLMOps: deployment and evaluation of LLMs, prompt versioning, guardrails
• Streaming: Kafka / Kinesis
• Spark / Data processing at scale

What Success Looks Like in 6 Months

• ML training and deployment pipelines are automated and standardized
• Model monitoring and drift detection is live for all production models
• Inference cost/latency is optimized and documented