zmievsa@gmail.com linkedin.com/in/zmievsa github.com/zmievsa
  • Created a product-oriented engineering culture in which engineers contribute to product decisions and focus on product outcomes
  • Mentored engineering and product teams on effective approaches to evals, agents, infrastructure, and LLMs
  • Established observability and DevOps practices, increasing release frequency from once every two weeks to 5–8 times a day
  • Led the hiring of strong heads of product, IT, and design
  • Established hiring (pipeline software and interview stages designed from scratch with clear, delegable criteria that preserve hiring quality), IT (SOC 2, device management, security), and HR (a workable system with extensive automation and proper vacation tracking) processes
  • Restructured and grew the engineering department from 3 to 15 engineers
  • Designed and developed “Cadwyn”, an open source API Versioning framework that reduced feature delivery time by roughly 40% and cut annual API management costs by more than €120,000
  • Proposed and built the infrastructure, processes, and culture for Continuous Delivery, speeding up feature releases from once every two weeks to multiple times a day
  • Led the Architectural Council, collaborating with the CTO and team leads to make company-wide technological and architectural decisions, ensuring alignment with business goals and technical excellence of Monite
  • Led the development and maintenance of all Python infrastructure: 54 internal libraries, 13 static analysis tools, 42 microservices, and all coding standards in the company
  • Mentored a senior DevOps engineer for seven months, resulting in his promotion to the team lead of DevOps platform
  • Designed and led the development of MoniteScript: a declarative DSL that allows the creation of complex business logic scenarios in distributed systems such as payable approval policies
  • Migrated the entire codebase (600,000+ lines of code) from monolithic schemas to a modular, Poetry-based package structure, resulting in a 90% reduction of cross-team merge conflicts and significantly decoupling the microservice architecture
  • Designed the highly scalable distributed “at least once” event/webhook gateway that supported signing more than $340,000 worth of client contracts
  • Reviewed all significant changes in the company’s API as a key member of the API council
  • Developed 2 JupyterLab extensions, 2 Python libraries, and a suite of syntactic-sugar modules that measurably accelerated the day-to-day workflows of the machine learning and data analysis teams
  • Overhauled the core of the project, enabling 9 core developers and over 200 client developers to utilize debuggers, static analysis tools, and coverage tools and transforming the long-term feature plans of the team
  • Mentored 2 developers for 3 months, resulting in a promotion for one of them after a quarterly review
  • Automated business processes, saving more than 700 work hours annually when processing documents
  • Built custom pipelines for company’s financial and geolocational data, eliminating 90% of time dedicated to data manipulation and saving the company more than $11,000 a year as a result
  • Developed and delivered the support infrastructure that connected users to administrators through a chat interface in social networks, saving $13,000 worth of annual moderation time for dealing with support tickets
  • Implemented CI/CD with automated building, linting, unit-testing, and deployment from scratch
  • Redesigned guidelines and development practices, speeding up the onboarding process for new developers by roughly 70%
 How Hiring Works at AI Companies Python community round table They are not unit tests: a survey of unit-testing anti-patterns They are not unit tests: a survey of unit-testing anti-patterns Death by a Thousand API Versions Co-organized Death by a Thousand API Versions API Change Management and Cadwyn Lessons on webhooks, platform engineering and API versioning API Design and Versioning Versioning Web APIs in Python The inner workings of large open-source projects
  • A Modern Stripe-like API Versioning framework
  • Provides an ability to support hundreds of API versions at the same time and backporting features from the latest version to any older version while keeping the business logic simple
  • Is used by dozens of companies and projects, including execution API in Airflow 3.0
  • Has more than 400,000 downloads per day from PyPI
  • Contributed to Python programming language by adding pathlib.Path.walk and speeding up os.walk by 18%
  • Made multiple contributions to both the discussion and the implementation of Type Intersections PEP
  • Automatic and lazy runtime generation of pydantic models that lowers the total size of schemas by up to 67%
  • Factory fixtures as classes that work well with dependency injection and language servers
 Python, FastAPI, PostgreSQL, Linux, OpenTelemetry, Docker, Redis, Kubernetes Javascript, TypeScript, Golang Hiring systems, org design, SOC 2, DevOps culture, product management