Skip to content
View bockuden's full-sized avatar

Block or report bockuden

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
bockuden/README.md

Hi, I'm Yehor Maksimov

Senior Software Engineer building resilient automation, developer tools, and reproducible data and ML systems.

15+ years in C# · Python and FastAPI · Browser Automation · Applied ML · Time-Series Research

I combine production-oriented software engineering with applied research. My work focuses on systems that remain testable, explainable, and reproducible under real-world constraints.

Featured open-source work

An open-source, ready-to-run failure sandbox for Playwright, Selenium, scrapers, and HTTP automation workers.

It provides deterministic stateful scenarios for transient 503 responses, permanent failures, delays, login flows, pagination, duplicate records, DOM changes, cancellation, and checkpoint/recovery testing. Runs are isolated by run_id, making failure sequences reproducible locally and in CI.

Published as a Python package with a CLI, container releases, a stable OpenAPI contract, and automated compatibility checks.

Repository · PyPI · Resilience Challenge

A production-style .NET and Playwright worker designed for interrupted jobs, repeated delivery, retries, DOM changes, duplicates, cancellation, and recovery from durable checkpoints.

The project includes SQLite persistence, bounded concurrency, structured logging, failure evidence, deterministic end-to-end tests, Docker Compose, and GitHub Actions.

It is validated against the independently released Resilient Automation Test Stand.

A reproducible balancing-authority-level sensitivity study of hypothetical large compute loads on the U.S. electrical grid.

The project emphasizes evidence provenance, data-quality auditing, explicit scenario boundaries, robustness checks, and reproducible offline builds.

A public-data research pipeline progressing from a state-level baseline to metro fixed-effects models and household-panel analysis.

The repository documents how data limitations, geographic resolution, and identification assumptions affect the conclusions.

Reproducible quantitative research on rolling correlation, Engle–Granger cointegration, stability, and out-of-sample mean-reversion tests for crypto perpetual pairs.

The study reports negative and inconclusive results alongside promising ones, without presenting exploratory backtests as production trading claims.

Background

  • 15+ years of C# development, browser automation, and web data extraction
  • Python, FastAPI, OpenCV, TensorFlow, and applied machine learning
  • Quantitative strategy research and backtesting
  • Time-series analysis and reproducible data pipelines
  • University degree in physics and mathematics education

Current focus

  • Open-source developer tooling
  • Resilient browser automation
  • Production-oriented data and ML pipelines
  • Reproducible quantitative research

Pinned Loading

  1. resilient-automation-test-stand resilient-automation-test-stand Public

    Deterministic failure scenarios for resilient browser automation testing.

    Python

  2. resilient-browser-automation resilient-browser-automation Public

    Production-style C# browser automation worker built with .NET 10 and Playwright, featuring retries, checkpoint/resume, idempotent processing, observability, and a Dockerized FastAPI test stand.

    C# 1

  3. openai-30gw-grid-exposure openai-30gw-grid-exposure Public

    Reproducible BA-level sensitivity analysis of hypothetical allocations of OpenAI's 30 GW compute target across the U.S. power grid.

    Python

  4. us-fertility-housing-income us-fertility-housing-income Public

    U.S. fertility and housing research pipeline: state-year MVP, metro-year FE upgrade, and SIPP household-panel V3.

    Python

  5. crypto-cointegration-research crypto-cointegration-research Public

    Reproducible research on rolling correlation, Engle-Granger cointegration, stability, and out-of-sample mean-reversion tests for Binance USDT-M perpetual pairs.

    Python