"Artificial Intelligence/Machine Learning Engineer"

Role: Artificial Intelligence/Machine Learning Engineer Location: Austin, TX (Hybrid) Functional Responsibilities: ERS is seeking a Machine Learning / AI Engineer with 12+ years of senior production experience and delivers AI-driven data reconciliation and analytics pipeline solutions in regulated environments. The Worker will design, build, and maintain the AI automation layer for the RISE data migration program, developing auditable anomaly detection pipelines, exception classification workflows, and real-time quality dashboards that accelerate conversion specialist throughput and provide ERS program leadership with continuous visibility into migration integrity. The worker will be responsible for: • Design and deploy ML-based anomaly detection pipelines layered on the Landing Zone to CDR ETL process, providing early-cycle flagging of data discrepancies before they propagate downstream • Build AI-assisted field mapping and classification tooling to accelerate source-to-target schema mapping across CDR cycles, enabling conversion specialists to apply prior resolution decisions consistently across subsequent cycles • Develop automated data quality scoring pipelines producing per-table and per-CDR-cycle quality metrics, providing QA and program leadership with real-time visibility into migration health • Apply LLM evaluation methodology and judge-model scoring frameworks to assess and validate AI-assisted reconciliation outputs for accuracy, consistency, and auditability • Develop and maintain lightweight, maintainable AI tooling that ERS-embedded staff can understand, operate, and extend following the engagement • Produce technical documentation of AI pipeline logic, model behavior, and automation design decisions in formats accessible to conversion specialists and program management • Actively participate in knowledge transfer sessions, helping ERS staff develop literacy in how AI was applied to the migration and what it produced The Worker should have deep production experience delivering AI-driven data reconciliation frameworks on Azure platforms, with demonstrated ability to build auditable anomaly detection and exception classification pipelines at scale, manage model performance in regulated environments (SOX, PCI-DSS, HIPAA), and communicate findings clearly to finance, actuarial, risk, and program leadership stakeholders. WORKER SKILLS AND QUALIFICATIONS Years Skills/Experience 6+ years' experience in Applied AI/ML pipeline development and deployment for large-scale data reconciliation programs; production experience building anomaly-detection, root-cause analysis, and exception classification models using PyTorch, Scikit-learn, and Azure Machine Learning in regulated financial or government environments 6+ years' experience Azure data platform engineering including Azure Databricks, Azure Data Factory, Azure Synapse Analytics, and Delta Lake; demonstrated ability to design automated, auditable reconciliation workflows eliminating manual row- and aggregate-level validation across multi-terabyte datasets 10+ years' experience Advanced T-SQL and PL/SQL development across SQL Server and Oracle including stored procedures, partition switching, column store indexing, and query optimization sustaining sub-second query response for high-volume ETL and dashboard workloads 6+ years' experience Rule-based exception classification pipelines and prioritized work queue construction; experience translating 30+ stakeholder control scenarios (finance, actuarial, risk) into automated validation logic, acceptance criteria, and agile backlog items 4+ years' experience Cloud-native ingestion pipeline engineering with Azure Data Factory, Azure Service Bus, and Azure Functions; schema validation, data lineage management with Azure Purview, and containerized micro-service deployment via Docker, AKS, and Git-based CI/CD 4+ years' experience Production model monitoring and drift detection using Azure Monitor metrics and custom drift detectors; MLflow experiment tracking and gradient-boosting ensemble tuning ensuring validation models retain statistical power across evolving data volumes and product mixes.