Senior Machine Learning Engineer
Core
Build and operate an LLM control plane, gateway, and agentic systems for an automotive retail platform, while maintaining classical ML pipelines for dealership operations.
Role type
Senior Machine Learning Engineer (LLM Platform & Agentic Systems)
Builds
LLM control plane, agent runtime, unified API/SDKs, and classical ML training/scoring pipelines for automotive retail.
Domain
Automotive retail technology, AI/LLM infrastructure, and data engineering.
Deliverable
production ML models | product features | infrastructure
Required skills
Python, Java/Scala/Go, distributed systems, microservices, MLOps, cloud engineering, LLM gateway design, agentic systems orchestration, graph databases, vector search, hybrid retrieval, model monitoring, data drift detection.
Preferred skills
AWS, Kubernetes, Docker, Neo4j, GraphQL, Spark/Flink/Kafka, MLflow, Airflow/Kubeflow, XGBoost/LightGBM, CI/CD for models, A/B testing, shadow/canary deployments.
Technologies
Python, Java, Scala, Go, AWS, Docker, Kubernetes, Neo4j, Neptune, TigerGraph, GraphQL, pgvector, Qdrant, Milvus, Airflow, Kubeflow, MLflow, Spark, Flink, Kafka, XGBoost, LightGBM
Responsibilities
Build LLM control plane/gateway with smart routing, rate limits, and cost tracking; ship unified API/SDKs with observability; enforce safety/privacy via content filtering and PII redaction; enable multi-vendor LLM usage with automated canarying; own agent runtime including tool registry and function calling; design orchestration patterns and manage agent state; enable classical ML training/scoring pipelines; monitor model/data drift and manage retraining; implement human-in-the-loop review; evolve domain graphs and entity resolution; serve real-time context with access controls; power hybrid search retrieval; run continuous evaluations for quality and safety; define SLOs for latency, uptime, and cost; maintain model registry and audit trails; provide developer templates and mentor teams on MLOps.
Seniority
Senior, hands-on IC with platform ownership
