About me

There is no such thing as a perfect tool to build a robust system. Rather, there is a tool (or tools) that is the most suitable for each particular job. And that’s what I do: I make sure that the best possible stack to solve this particular problem is used. That the codebase is clear, simple and maintainable. That documentation is present AND USEFUL!

I constantly learn new things at work and in my free time: I like combining new and time-tested approaches to solve a problem I’m working on. I participate in conferences and workshops, sometimes as a listener and sometimes as a presenter: I enjoy giving back to the community that taught me so many useful skills for free.

Programming languages

• Python: 8+ years
• Java/Kotlin: 3+ years
• Rust: 2+ year
• JavaScript: 2+ years
• R: 2 years
• Lua: 2+ year
• Dart: 2+ years
• C#: 1 year

Data Bases

• Postgres, Snowflake, MySQL, MariaDB, BigQuery, SQLlite, MS Server
• Elasticsearch
• Redis
• Blob storages
• Neo4j
• Hive/Presto
• Cassandra

AI/ML/DS

• Classic ML: supervised and un-supervised models
• Deep learning: NNs, LSTMs, Transformers
• NLP: LLMs, classic ML
• Frameworks: langchain (ml-flow), opencv, tensorflow, ragas
• Custom data quality assurance pipelines: data drift detection, anomalies handling, fraud detection
• 3d parties LLMs: OpenAI, Mistral, open-source, Anthropic, Open-Router

Frameworks

• Flask, FastAPI, HttpIO, Pytest, Django, SQLAlchemy, Celery, RQ
• Dropwizard, Spring, Mock, Hystrix+, FlyWay
• Axum, Rocket, Actix-web, Sqlx, Diesel, GGEZ, Amethyst
• Node.js, React, JQuery, Angular
• Flutter
• ASP.Zero, EF, ASP.Core
• Love2d

DevOps/Administration

• Docker, Podman
• Ansible
• Terraform
• Linux
• K8s
• CI/CD (Heroku, Gitlab, Bamboo, Jenkins, GitHubActions)
• AWS, Google Cloud, Hetzner Cloud, Azure
• Prometheus, Grafana, ELK, Jaeger, Sentry, DataDog
• Bash, Makefile, Justfile

Misc

• RabbitMQ, Kafka
• REST, GraphQL
• Hyperfine, Oha, Heaptrack, Valgrind, PyFlame
• Selenium, Cypress, Playwright
• OOP, ECS (entity component system)
• Natural languages: Russian (native speaker), German (fluent), English (fluent)
• Data Analysis: Metabase, Redash, Tableau, Holisitcs
• Graphic design/Prototyping: Inkspace, PixelArt, Tiled, Figma

Projects

• Json configurable fully connected NN
• Python vs R on a simple statistical example
• Simple CPU simulation with MyHDL
• Hyper Oktagon - gameplay
• Hyper Oktagon - code
• Personal Tracking REST API
• Web frameworks with PG on the back comparison
• (Private): Personal Finance analytics and tracking, mental tracking tool - Deployed on Hetzner Cloud server behind ngnix/caddy, utilizes ansible for easier administration of the server and containers
• Ask me about my private projects
• (Private for now): Positive training: LLM generated heavy life scenatio to react to. Flutter for user-interaction, Rust/Python/LLMs APIs on the back with a simple custom quality monitoring system

Talks

Tipps and tricks to improve stress resiliance

Pyladies HH: Intro to Docker

PP2020: How to deal with a burnout during quarantine

BigData BBQ 2019: What features matter

Pyladies/PyData: How to use rust to speed up your python (PyO3 short talk)

EP2019: Python vs Rust for simulation

AMLD2019: Data analysis workshop organizer

Pyladies 2018.12: Game development in python (hands-on workshop)

EP2018: Data is not flat

EP2017: Baby steps in short-text classification

EP2023: Quantify self

DevDays2024: Scientific approach to Flow

Professional Experience

Freelance (2022-02 – ...)

Software and Infrastructure consultant

aestimium GmbH (2020-12 – 2022-01)

Senior Machine Learning Engineer (Bakend, Data pipelines, BI and Reporting, ML/DS)

FREE NOW (formerly MyTaxi) (2018-08 – 2020-10)

Machine Learning Engineer (Data preparation; streaming; chron jobs; backend engineering; highly specific, high performance simulation framework)

Joblift (2016-04 – 2018-05)

Fulltime software developer (Data Engineering, Backend Development, NLP)

Eversolve (2015-08 – 2016-01)

Working student, front end developer (Angular.js, Bootstrap, JS)

Everbase AG (2015-05 – 2015-08)

Working student, Graphics manipulation, design

Education

Gymnasium №3, Novosibirsk, Russian Federation(1997-2007)

Novosibirsk State University, Economical Department, Economics (2007-2011)

The bachelor thesis consisted in proving that classical portfolio models are ineffective on modern Russian stock exchanges. Thus I suggested an adopted approach for choosing the portfolio for a private investor.

University Hamburg, MIN faculty, Information Systems study (2011-2015)

The goal of my second bachelor thesis was to check what language is more suitable for an econometric task written on the beginner level. Python and R were chosen for comparison, since these two languages are heavily used by both scientific and non-scientific communities. To answer this question I have written two identical Programs (with respect to the language's features and idiomatic style) in Python and R and have run performace and memory usage tests. I have also compared subjective usability (ease of use, documentation quality, community) of the langages.