Tomáš Drtina

About Me

Making sure our AI/ML models are connected with our FE apps. Everything from BE (servers, DBs, containers, etc.) to CI/CD and other DevOps, from architecture to development. Passionate about making everything secure.

Rust language and networking have a special place in my heart.

Work Experience

Software Architect, DevOps Engineer, Tech Lead

Blindspot SolutionsJuly 2015 - Present

In Blindspot my main focus is on developing backend services. I cooperate with frontend teams to design and implement APIs and secure them, but also with AI/ML teams to connect backend service to their models.

I'm also managing internal servers and our cloud services as a DevOps engineer/administartor.

Work Projects

Improvement of client's internal CI/CD pipelines and developer productivity

April 2020 - Now

Created new integration tests in NixOS using QEMU VM
Managed and extend Jenkins jobs
Optimized Docker images
Improved speed, cacheability and robustness of tests

Role:: DevOps engineer
Technologies:: NixOS, systemd, Jenkins, Bash, GreenplumDB, Python, Scala, Go

Custom chatbot application

March 2022 - Now

Created multi-tenant application based on Google Cloud Functions.
Created REST and MQ-based API for GPU inference microservices.
Deployed chat microservices to Google Kubernetes Engine with GPU nodes.

Role:: Architect, DevOps, ML developer
Technologies:: Python, Kubernetes, Terraform, TypeScript/Node.js
Cloud technologies:: Google Kubernetes Engine, Google Filestore, Google Cloud Functions, Firebase Authentication, Firebase Realtime Database

Customization of ARIS platform for a custom client deployment

July 2021 - January 2022

Customized Blindspot's ARIS platform for a client's customer support center.
Deployed to customer's on-premise servers
Integraed with Microsoft 365 Graph API and customer's RPA processes.

Role:: Backend developer, DevOps engineer
Technologies:: Kotlin, Spring Boot, Docker, PostgreSQL, Python, Scala

Custom AI image processing app

February 2021 - May 2021

Created a Node.js/TypeScript-based API deployed as Azure Web Service.
Deployed AI models as Python microservices, including GPU-based neural network service.
Deployed to Azure Cloud using custom-built VM image built with Packer.
Created a job orchestration pipeline using Azure Service Bus.

Role:: Architect, backend developer, DevOps engineer
Technologies:: TypeScript, Node.js, Packer, Python, PostgreSQL, Azure Service Bus

Logistics processes optimization

August 2019 - April 2020

Built an optimization tool for logistics operators to improve the onboarding of new employees and new processes.
Built on Blindspot's Optimus platform
Deployed to a hybrir environment on Azure Cloud and on-prem Azure Stack.
Integrated with client's AD using SAML and using mutual-TLS for all communication.

Role:: Tech lead, architect, lead backend developer, DevOps/deployment/security
Technologies:: Kotlin, Spring Boot, JPA/Hibernate, PostgreSQL, Docker, Terraform
Cloud technologies:: Azure Cloud (Azure Web Services, Azure Database for PostgreSQL, Azure Blob Storage, Azure DataLake Storage, Azure API Management), Azure Stack (Azure Web Services)