全职 Machine Learning Engineer (Ops) 工作, 薪水, Newbridge 公司招聘中 - Ricebowl

Role Overview

Own the intersection of Data Governance and MLOps. You will build trusted data and trusted AI - ensuring our clients data is accurate, compliant, and governed, and our clients ML models are reproducible, monitored, and responsibly deployed to production.

This role is 50% Data Governance, 50% MLOps / ML Platform Governance.

Key Responsibilities

A. Data Governance (50%)

1. Framework & Stewardship

• Design and run enterprise Data Governance framework, policies, and RACI for data owners/stewards
• Establish Data Governance Council and operating model across Product, Engineering, Analytics, and Business
• Define KPIs: catalog coverage, data quality score, policy adherence

Data Quality, Catalog & Lineage

• Implement business glossary, data catalog (Collibra / Alation / Purview / DataHub), and end-to-end lineage
• Define and monitor data quality rules, SLAs, anomaly detection for critical domains (Customer, Product, Transaction)
• Manage data classification, PII/PHI tagging, retention, and access control policies

Compliance & Security

• Ensure compliance with PDPA, GDPR, CCPA and internal security standards
• Partner with DPO / Legal / GRC for consent, purpose limitation, anonymization, and audit readiness
• Own access governance - RBAC/ABAC for data warehouse, lakehouse, and feature store

MLOps & AI Governance (50%)

ML Lifecycle & Platform

• Own MLOps best practices: from feature engineering -> training -> validation -> deployment -> monitoring
• Build and manage ML platform components: Feature Store (Feast / Tecton / SageMaker Feature Store), Model Registry (MLflow / SageMaker Model Registry), Experiment Tracking
• Standardize CI/CD/CT for ML with Git, Docker, Airflow / Kubeflow / SageMaker Pipelines

Model Governance & Responsible AI

• Implement Model Governance: model inventory, model cards, lineage (data -> features -> model -> endpoint), approval workflows
• Enforce responsible AI checks: bias/fairness, explainability, drift, and reproducibility
• Align with AI Governance frameworks: NIST AI RMF, Singapore Model AI Governance Framework, AI Verify, ISO 42001

Monitoring & Operations

• Implement monitoring for data drift, concept drift, feature skew, and model performance degradation
• Set up alerting, automated retraining triggers, and rollback strategies
• Optimize model serving costs, latency, and scalability on AWS / Azure / GCP

Tech Stack You Will Work With

Governance: Collibra, Alation, Purview, Informatica, DataHub, AWS Glue, Apache Atlas

Data: Snowflake / BigQuery / Redshift, S3 / GCS, dbt, Airflow, Spark, Kafka

MLOps: MLflow, Kubeflow, SageMaker, Vertex AI, Feast, Evidently, Great Expectations, Docker, Kubernetes, GitHub Actions

Languages: Python (must), SQL (must), PySpark

Requirements

• 6-10 years total in Data Engineering / Data Governance / MLOps
• At least 2+ years owning data governance and at least 2+ years deploying ML models to production
• Strong hands-on with DAMA-DMBOK and MLOps principles
• Proven experience setting up Model Registry, Feature Store, and monitoring for production ML systems
• Deep understanding of PDPA/GDPR, data security, and AI risk
• Excellent stakeholder management - you can talk to both Data Scientists and Risk/Legal

Nice-to-Have

• CDMP, AWS Certified ML Specialty, or similar
• Experience with LLM / GenAI governance - prompt logging, RAG governance, hallucination monitoring
• Experience with Great Expectations, Monte Carlo, Evidently AI
• Industry experience in Media, FinTech, or other regulated industry

What Success Looks Like in 12 Months

• Top 5 data domains governed with SLAs and quality monitoring >95%
• 100% of production models registered with model cards, lineage, and approval workflow
• Automated drift detection live for all critical models with <2hr alert SLA
• Data catalog adoption >80% and zero compliance audit findings