Igor Światowski
Devops Engineer @Deloitte
Signup · Get unlimited contacts
WORK HISTORY
Devops Engineer @Deloitte
Build and run the infrastructure that ships ML models to production for a major European bank. Training pipelines, model registry, serving endpoints, monitoring - responsible for the full chain, from training pipelines through to model serving in production.Architected the MLOps platform on AWS - SageMaker for training and inference, MLflow for experiment tracking and model versioning, ECR for container registry, all provisioned through Terraform and deployed via GitHub Actions. Designed the CI/CD path so data scientists push code and models reach production without manual intervention - automated validation, containerization, canary deployments, and rollback. Built model monitoring and drift detection pipelines that catch degradation before it hits downstream systems. Run EKS clusters serving both ML inference workloads and supporting microservices - namespace isolation, autoscaling, ingress, and GPU node scheduling. All infrastructure managed as code through Terraform with no manual provisioning.
EDUCATION
WSB Schools of Banking in Poland
Bachelor of Science, Computer Science
ABOUT IGOR ŚWIATOWSKI
DevOps / Platform / MLOps Engineer focused on building scalable, production-grade infrastructure on AWS and GCP. I design and ship end-to-end pipelines - from application deployments to ML model serving - using Terraform, Kubernetes, and modern CI/CD tooling. What I do day-to-day: → Design and automate cloud infrastructure as code on AWS using Terraform, CloudFormation, and Pulumi → Build and manage Kubernetes clusters (EKS) to run microservices and ML workloads at scale → Architect CI/CD pipelines with Jenkins, GitHub Actions, and ArgoCD for both application code and ML model delivery → Ship ML models to production — containerization, model registry integration, inference endpoint automation, and monitoring → Implement MLOps practices: automated training pipelines, model versioning, A/B testing, drift detection, and observability with SageMaker, MLflow, and Kubeflow → Provide stable infrastructure for applications, APIs, and data pipelines while keeping costs under control → Automate everything with Bash and Python — if it runs more than twice, it gets scripted Tech I work with: AWS (EKS, ECS, Lambda, SageMaker, S3, ECR, CloudWatch, IAM), GCP, Terraform, Kubernetes, Docker, Helm, Jenkins, GitHub Actions, ArgoCD, Linux, Git, Ansible, Prometheus, Grafana, MLflow, Kubeflow, Python, Bash Currently exploring deeper specialization in MLOps — bridging the gap between data science teams and production infrastructure. I believe the next wave of DevOps is making ML delivery as reliable and repeatable as traditional software delivery. Open to interesting roles, collaborations, and conversations. Let\'s connect.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.