MLOps platforms
End-to-end platforms for experiment tracking, model registries, training, deployment, and continuous retraining. A consistent path from research to production.
MLflow · SageMaker · Azure ML
MLOps / LLMOps Platform Engineer
AI into production.
Reliability at scale.
I build the platforms that bring machine learning and generative AI into production—and keep them running reliably.
I'm Koby, a platform engineer with 9+ years of experience across ML infrastructure, automation, and observability. Currently at State Farm, I support production AI systems for insurance. Previously at JPMorgan Chase and Charter Communications.
Miami, Florida Cloud-native. Production-focused.
9+years in data & ML infrastructure
50+production ML models supported at State Farm
60%+reduction in release cycle time at State Farm
Expertise
What I do
From the first experiment to the systems
that serve it in production.
LLMOps & generative AI
Infrastructure for RAG applications and AI agents, with prompt versioning, evaluation harnesses, inference serving, and visibility into cost and latency.
RAG · LangGraph · vLLM · Bedrock
Cloud & Kubernetes
Kubernetes-native infrastructure for CPU and GPU workloads. Reusable infrastructure modules, autoscaling, and team isolation built for enterprise environments.
EKS · AKS · Terraform · Helm
GitOps & delivery automation
Version-controlled model delivery with automated training, evaluation, and promotion. Repeatable releases, audit trails, and dependable rollback workflows.
Argo CD · Harness · GitHub Actions
AIOps & observability
Model performance, data drift, and infrastructure telemetry brought together. Monitoring and automated alerting that help teams catch degradation early.
Prometheus · Grafana · PagerDuty
Data & feature engineering
Scalable data pipelines, production-ready feature sets, and quality checks that protect downstream analytics and model training workflows.
Python · SQL · Snowflake · Kafka
Career
Experience
Building production systems across
insurance, banking, and telecommunications.
MLOps / LLMOps Platform Engineer
Architect and operate a centralized MLOps platform supporting 50+ production models across pricing, claims automation, fraud detection, and telematics.
- Reduced release cycle time by over 60% through automated model training, evaluation, packaging, and promotion.
- Built GitOps delivery with Argo CD and Harness, including audit trails and rollback across environments.
- Designed LLMOps infrastructure for inference, prompt versioning, RAG orchestration, and evaluation.
- Unified model, drift, and infrastructure monitoring while enabling self-service ML infrastructure through Terraform and Helm.
MLOps Platform Engineer
Delivered production ML and generative AI systems for private banking, from prototyping through live deployment, ongoing operations, and continuous retraining.
- Built containerized model delivery, drift monitoring, retraining triggers, and SLA dashboards across AWS SageMaker and Azure ML.
- Engineered RAG pipelines with prompt management, evaluation frameworks, and enterprise knowledge integration.
- Standardized infrastructure and model delivery in regulated environments, and mentored engineers through code reviews and internal playbooks.
DataOps Engineer
Built Python-based data pipelines at multi-terabyte scale, creating reliable foundations for analytics and production machine learning.
- Designed validation frameworks to catch schema drift, null violations, and statistical anomalies before downstream use.
- Automated data pipeline deployments with Jenkins, bash, and containerized workloads.
- Engineered features for demand forecasting and customer churn models using pandas and SQL.
Learning & development
Education & credentials
Education
Bachelor of Science
in Psychology
Florida International University
2010Certification
Professional
Scrum Master
PSM
Professional development
AWS Certified AI Cloud
Practitioner Foundational
Certification studies
In progressGet in touch
Let's build what comes next.
For conversations about MLOps, LLMOps,
and reliable AI infrastructure.