Albert Levin

Leading architecture and delivery of a multi-tenant enterprise AI and ML platform for internal developer and business teams.

• Designed and built a multi-tenant AI and ML platform used by 50+ developers and 200+ business users, covering data, model development, deployment, AI applications, and platform operations.
• Defined platform architecture, technical standards, and delivery requirements across internal engineering teams, business stakeholders, and external technology partners.
• Built reusable Infrastructure as Code and CI/CD components with Terraform and Azure DevOps, accelerating onboarding and standardizing deployments across environments.
• Implemented enterprise AI solutions from self-hosted LLM coding agents and agentic applications to conventional ML workloads on Kubernetes and Databricks.
• Established platform governance, security, and observability with centralized AI gateway controls and end-to-end runtime visibility.
• Translated GDPR and EU AI Act requirements into practical platform controls, deployment processes, and infrastructure standards.

Designed and implemented MLOps workflows and production AI applications for enterprise clients.

• Designed and implemented MLOps workflows with Databricks and MLflow, including model registration, lineage, reproducibility, and automated deployment.
• Built production AI applications and agents, including knowledge-base and natural-language-to-SQL solutions, using Python, TypeScript, LangChain, Next.js, and React.
• Developed and deployed ML model APIs on Kubernetes using Python, FastAPI, and Docker with production-oriented CI/CD and infrastructure automation.
• Built and operated self-hosted LLM inference workloads on Kubernetes using Hugging Face, Docker, and Helm.
• Designed and maintained cloud infrastructure with Terraform and Terragrunt, including Go-based Terratest automation.

Led data platform and data engineering delivery for ML-focused workloads.

• Designed and optimized ETL/ELT pipelines in Databricks using PySpark and Delta Lake for ML training, feature engineering, and data processing.
• Redesigned processing around an event-driven Azure architecture with Databricks, PySpark, and Event Hubs, reducing pipeline runtime from 2.5 hours to less than 1 hour.
• Implemented automated CI/CD and testing for Python packages in Azure DevOps, improving reliability and repeatability.
• Built custom monitoring and observability tooling using Python and Grafana.
• Developed a VS Code extension for local Python and PySpark job debugging, reducing troubleshooting cycles across the team.

Developed and productionized ML models and model-serving systems for retail use cases.

• Developed ML models for forecasting, fraud detection, recommendation, and search using XGBoost, scikit-learn, and TensorFlow/Keras.
• Built ML model-serving APIs with Python, FastAPI, Docker, and Kubernetes on GKE and AKS, with production monitoring via Prometheus and Grafana.
• Implemented MLOps workflows using Azure DevOps, Databricks, and MLflow for reproducible model development and deployment.
• Developed PySpark and Databricks pipelines for ML training, feature engineering, and feature-store workloads.
• Automated deployment and operations with Kubernetes, Helm, Docker, and Terraform.

Developed data integration workflows and reporting pipelines for European operations.

• Developed ETL/ELT pipelines with Python, Airflow, and BigQuery.
• Developed internal web applications with Django and Flask.
• Designed data models in the enterprise DWH using Data Vault 2.0.
• Set up CI/CD pipelines in GitLab and managed an Airflow Kubernetes cluster.
• Managed cloud network and security controls in Google Cloud, including firewalls, VPCs, and load balancers.