Skip to main content



My multi-year span in the old-fashioned insurance industry has proven to be surprisingly versatile and interesting. I suppose even in a centuries old industry you can get to work on interesting tech stack, as long as you position yourself properly under innovative, brave managers and among the producers, not the support stuff and can use technology for the benefit of others, especially less tech-savvy users. The fact that the industry did not adopt the cloud "revolution" is a blessing in disguise - many of my most interesting activities would have been outsourced by cloud providers (which is a constant threat to e.g. system admins in any industry).

So in my Data Scientist / ML Engineer hat I work in python on fully automated ML python modeling pipelines (Papermill+Scrapbook, MLflow), from data munging, feature engineering and selection (incl. SHAP), training, distributed hyperparameters tuning (Optuna), reproducibility and automated validation (MLflow) and automated periodic monitoring of post-production model performance (Papermill, MLflow, CronJobs).

I've been responsible for building and productionalizing first successful ML models in both major areas of our business (demand and risk models) and a paradigm shift away from decades-old (generalized) linear models that still dominate in the industry.

In my MLOps hat I develop in-house docker containers (allowing for self-service package installations and automated updates and security scans) for data scientists working on analyses and ML models development (GPU-enabled, Python, R, H2O) with familiar interfaces such as Jupyter Notebook/Lab, RStudio Server, and VS Code, specialized ML Ops frameworks such as MLFlow or generic data and file management tools / in-house data lakes (such as S3/MinIO and Cloud Commander), and SQL and no-SQL databases (notably MariaDB and Redis). I also develop and maintain in production and staging clusters custom apps with REST APIs for the production deployment of ML models and their features (using Python or Java, Flask/FastAPI, gunicorn, Redis, MinIO, git, and bash).

In my DevOps hat I orchestrate two types of ML containers (stateful for ML models development and stateless for their production deployment) in an array of on-prem air-gapped k8s/OKD clusters, automating multi-layer container builds, packages/libraries/extensions updates, security scans, and staging/production deployments using Jenkins pipelines, bash, python, and the dreaded Groovy scripts and build and deployment configs and templates, integrated using webhooks.

I also perform linux systems admin role for CI/CD build servers (CentOS/Stream, docker/compose, Jenkins, NGINX, Postgres, Clair, MicroK8s) and fulfill an Openshift/OKD 3/4 business admin roles for multiple clusters (using OCP/OKD/kubectl CLI, YAML configs and bash scripts) in both the data science development and in ML models production clusters for dozens of company data scientists and tens of ML models in production.

Top Answers