# Gennadii Nikitin

QA Engineer (Senior) · 8+ years · Manual + Automation + API

LATAM, Asunción, Paraguay · Remote (UTC-3), overlapping CET and EST · relocation considered · available immediately

Contract: B2B, EOR or direct contract — negotiable; also self-employed for RU entities.

[g.nikitin14@gmail.com](mailto:g.nikitin14@gmail.com) · [Telegram @ni_genn](https://t.me/ni_genn) · [LinkedIn](https://www.linkedin.com/in/gennadiy-nikitin-21337022b) · [GitHub](https://github.com/ge-a-nikitin)

---

## Summary

QA engineer with 8+ years on desktop, web and mobile client-server products in high-traffic consumer promos and gamification — prize games, seasonal campaigns, loyalty mechanics — including a prize promotion with up to 3M+ unique monthly users. I own the full SDLC on my projects: requirements and testability before development, test design, functional / regression / API / integration testing, and automation where it pays off — TypeScript on the company’s in-house Playwright-based framework, k6 load testing on two projects. I leave documentation and processes a team can pick up and reuse.

## Key achievements

- **2 days → 2 h** — Regression cut from 2 working days for 2 testers to ~2 hours for 1 tester by automating core desktop scenarios with Python + PyAutoGUI and image-recognition locators.
- **PM role** — Took over most of the project-manager role on Pushkin 225 Years without the title while the team was short-staffed — launch readiness, daily statuses, backlog priority, requirements with the stakeholder product, API and Kafka research, mentoring new colleagues. The promo launched on schedule with me as its only QA.
- **30%** — of L1 requests closed by an AI chatbot on prepared documentation, never reaching the team; tracked per channel and project on a dashboard.
- **5 + ~10** — projects on the department automation template I co-designed on eva.io: five of my own, about ten more in adjacent teams.
- **QA-owned** — performance testing: moved from production-only, backend-owned runs to a dedicated environment owned by QA — benchmark, stability and maxperf in k6 on two projects, metrics in Grafana and Cosmos; the approved methodology with template and quickstart is reused on new projects.
- **Release gate** — Automated suites gated every release where we wrote e2e tests: a red pipeline held it until we knew why.
- **Daily CI** — Visual regression (screenshot tests) on daily CI runs, reports delivered to the corporate messenger.

## Core skills

- **Testing:** Functional, regression, integration, system, UI, API, cross-browser, performance/load, visual regression (screenshots).
- **Test design & analysis:** Boundary value analysis, equivalence partitioning, state transition diagrams, requirements analysis, 3 Amigos sessions (PM, developer and QA together), testability review.
- **Test automation:** TypeScript on eva.io (Selenium at first, then Playwright; most work in the hybrid period), Playwright (eva.io modules at work, the public e2e suite below), Python (PyAutoGUI, image recognition, Tillypad), parameterized tests, visual regression, Allure TestOps reporting, reusable module and page structure.
- **API & data:** Postman, Insomnia, Swagger/OpenAPI, Charles, Fiddler, OWASP, SQL (SELECT, JOIN, GROUP BY, subqueries, INSERT/UPDATE), MSSQL replication verification.
- **Performance:** k6 (benchmark, stability, maxperf), load profiles built from production logs, Grafana (plus the internal metrics service, Cosmos).
- **Platforms & infrastructure:** Docker, BrowserStack, real iOS/Android devices, Android Studio, Chrome and Firefox DevTools.
- **AI in testing:** LLM-assisted checklists and test scenarios, prompt iteration, Nestor AI (the company’s own Cline-like agent) for test code in the IDE; shared the practice with the team.
- **QA process:** Test strategy, test plans and test cases, shift-left (requirements and testability review before development), release criteria and quality gates, defect lifecycle and triage, test documentation, QA metrics; Agile — Scrumban (Scrum/Kanban).
- **Tools & CI/CD:** Git, GitLab CI with daily scheduled runs, Jira, Confluence, Allure TestOps, Redmine, Kafka topic integrations.

## Experience

Current status: left T-Bank for personal reasons.

### Quality Engineer — T-Bank (ex Tinkoff)

April 2023 – September 2026

Web and webview applications inside Android/iOS apps: promo mechanics that grew into products — a monopoly-style prize game that reached up to 3M+ unique monthly users, longer-lived services, multiple third-party integrations. Sole QA for the first two years, then one of two equal QA engineers after the teams were rebuilt around larger products; owned the test strategy for my own projects. Work ran in Scrumban, the team’s name for a Scrum/Kanban hybrid chosen for project-based work. Release cycle: 3–7 months of development on staging, one release day, then follow-ups daily to weekly by severity. On-call for support and alerts: QA triages and routes to PM, developer or team lead; incidents reviewed with backend and the lead.

- **Automation across the department:** co-designed the department’s test template on eva.io (TypeScript; Selenium first, Playwright modules after the migration) — authorization via the internal test environment QA Utils (ready-made test users), test execution and Allure TestOps reporting, repository structure, deployment guide; reviewed code and contributed to the template repo; adopted across own and adjacent teams.
- **Game Center (unified entry point for special projects):** built the testing process from scratch — test plans, Allure reports, FAQ and knowledge base; ran system, UI, API and integration testing; added screenshot-based visual regression on the Playwright module in eva.io with CI/CD runs, red about once a month when seasonal assets or the project list changed; handed over with documentation and processes in place.
- **Test pyramid and TestOps:** with the second QA over the last year and a half — unit and integration runs from the backend and frontend repos, e2e and screenshot suites, all reporting into TestOps: statistics, test-case management, and unit, integration and e2e layers kept deliberately separate. A contract sweep and field validation across the API layer caught a broken API contract and a field left unvalidated.
- **Shift-left as the default, not an initiative:** in requirements before development started — co-wrote functional requirements in 3 Amigos sessions, worked to the department RFC on involving testing earlier, kept a running page of open questions during design, and called out weak spots while changing plans was still cheap.
- **Support load:** set up AI chatbot auto-responses in the corporate messenger from prepared documentation; a per-channel, per-project dashboard showed 30% of inquiries resolved there, never reaching the team; kept ongoing support and defect tracking.
- **Performance testing:** with SRE, set up a dedicated performance environment; benchmark, stability and maxperf scenarios in k6; metrics in Grafana and Cosmos. The runs exposed database saturation — slow queries and lock contention under load — and it was fixed before launch. Wrote the department-approved methodology plus a template and quickstart the QA team reused on new projects.
- **Pushkin 225 Years: sole QA and most of the PM role:** with the team short-staffed, took over the PM’s operational work — daily statuses and priority, launch readiness and dates with the lead PM, the launch content, requirements from the stakeholder product — and mentored the two new joiners (a backend developer on probation, an intern PM) through onboarding and 1:1s.
- **Cross-team relief:** twice covered a QA gap in a neighbouring team for 1–2 months: release testing plus the strategy context, releases kept moving with no resident QA, handed back with documentation.

### Quality Engineer — Tillypad

May 2018 – April 2023

Grew from Test Specialist through Test Engineer to Quality Engineer in a team of 3 testers and 6 developers. Client-server system for restaurant automation: desktop apps for waitstaff, bartenders, cooks and managers, with mobile (Android/iOS) and web versions added later; MSSQL with replication between the restaurant database and head office / cloud storage.

- **Full-cycle manual testing and support:** created and maintained test plans and test cases; executed unit, integration, system and regression testing; API checks in Postman and Swagger, traffic analysis in Charles and Fiddler; tracked defects in Redmine; elaborated requirements to catch weak spots early in the cycle; troubleshot user issues with the support team and onboarded new team members on the product and processes.
- **UI automation:** automated core desktop scenarios in Python + PyAutoGUI, using image recognition of buttons and UI elements for coordinate detection and baseline comparison; regression went from 2 working days (2 testers) to ~2 hours (1 tester).
- **MSSQL replication testing:** helped shape the approach; wrote the SQL scripts that insert, update and delete source data, then verified the target database held what should replicate and nothing else; ran at the start of every release cycle — around ten defects over three years, all high severity, on logic the product could not work without.

## Selected projects

- **hecho-demo** — HECHO, a client-server order board built to demonstrate QA practice end to end: React 19 + TypeScript SPA, NestJS API with Swagger/OpenAPI, PostgreSQL + Prisma, SeaweedFS, Docker Compose; Vitest unit and API integration tests, bilingual product and QA documentation, GitHub Issues bug-report template, CI deployment. [Details →](/demo/) · [GitHub](https://github.com/ge-a-nikitin/hecho-demo)
- **hecho-e2e** — Playwright end-to-end suite driving a real stack (web, API, PostgreSQL, SeaweedFS): customer-order cases across chromium and Pixel 7 projects, tags, Qase TestOps catalogue sync, SUT pinned by commit, reports. [GitHub](https://github.com/ge-a-nikitin/hecho-e2e)
- **TShop** — T-Bank’s merchandise store inside the mobile app; order-fulfilment integrations, sole QA.
- **Birthday 4.0** — congratulatory service with prize giveaways that grew from a one-off promo into a product; testing from scratch, automation on the framework, then two years as the knowledge holder and entry point for its testing questions — support requests, investigations, triage, release fixes.
- **New Year 2024** — annual gamified promotion; Kafka topic integrations and deep-link testing across app versions. Pulled the app versions users actually ran from the logging system into charts on the department wiki, and those numbers decided which versions we tested.
- **Pushkin 225 Years** — quests, prizes and loyalty for the poet’s 225th anniversary, built from scratch: development April – July 2024, live July – August 2024. Sole QA across frontend and backend — shift-left requirements, Kafka topic integrations for the new task types, API scripts for working with the project, AI-generated test data and helpers. Automation was a conscious no: timeline, staffing and functional depth left no room for a suite that would pay for itself.
- **Stickermania with Vkusno i Tochka** — co-promotion with the restaurant chain; parameterised automated tests, an authorization module added to the shared template, AI-assisted checklists. An R&D study on parameterised layout automation showed a high false-positive rate, so we dropped it and saved the resources; the parameterisation itself went into functional autotests, cutting duplication.
- **UserAgent Redirect** — browser user-agent research for a fallback page; findings in an ADR plus a usage guide, automated tests over large value lists manual checks could not cover, knowledge-sharing session for the department’s teams.
- **Prizopoly 2.0** — prize promotion reaching up to 3M+ unique monthly users; k6 performance testing covering every critical endpoint available at the time, three planned scenarios distilled into one universal one, load-testing ownership moved from backend to QA, results presented to both teams. [T-Bank press release](https://www.tbank.ru/about/news/21102024-t-bank-launched-prizopoly-an-interactive-game-with-the-largest-prize-pool-in-its-history)

## Education & languages

- St. Petersburg College of Technology (sptcol.ru) — Programming, secondary vocational education, 2023 – 2027 (expected).
- Emperor Alexander I St. Petersburg State Transport University (PGUPS) — Faculty of Mechanical Engineering, Railway Rolling Stock and Wagon Facilities, 2007 – 2011, three years completed (incomplete higher education).
- Training: internal employer courses before each new scope — end-to-end automation on eva.io and load testing with k6. No external certifications: SQL, test theory, programming and English come from college, university and self-study.
- Russian — native (C2). English — B2, working proficiency in reading, writing and speaking: documentation, tickets, reports and interviews.
