Pavel Lesyuk

QA Architect & SDET Lead

pavel@lesyuk.dev lesyuk.dev linkedin.com/in/lesyuk github.com/lesyuk Global Remote

Summary

QA Architect and SDET Lead with 8+ years building quality and security infrastructure for high-load consumer and Web3 products. Secured a Telegram marketplace that cleared $280M+ in volume and a platform serving 43M+ monthly active users, closing 30+ critical vulnerabilities before launch. Maintained a record of no known critical vulnerabilities reaching production within my area of responsibility. I build deterministic control planes for agentic AI, so LLM agents accelerate the testing lifecycle without ever authorizing their own gates. Equally at home threat modeling on-chain settlement, tuning backends to 50k RPS at sub-300ms, and shipping fixes directly into Go and React.

Core competencies

  • Agentic AI QAProcess-as-Code · Claude Code · Cursor · Codex · MCP integrations · Agentic Pipelines · Self-Repairing Test Automation
  • Web3 & SecurityThreat Modeling · TON Ecosystem · Telegram Mini Apps · On-Chain Settlement · Burp Suite · Frida · Fuzzing · gosec · trivy · govulncheck
  • Test ArchitectureQA Strategy & SDET Leadership · Playwright · Appium (iOS/Android) · k6 Load Testing · Contract & API Testing · Hardware-in-the-Loop · Allure TestOps
  • Infra & ObservabilityGitHub Actions · GitLab CI · TeamCity · Docker & Kubernetes · Grafana · ELK · Firebase
  • LanguagesTypeScript · Go · Python · SQL · Bash

Professional experience

QA Automation Architect — The Open Platform

Jan 2026 – Jul 2026 · Remote

  • Architected an AI-orchestrated QA system where Claude Code agents drive the testing lifecycle autonomously, with adversarial test design for agentic wallets against domain invariants — revoke-atomicity, wallet isolation, capped spend — behind mandatory human review gates.
  • Eliminated live-fund exposure from CI by engineering a deterministic Playwright pipeline over a mocked TON API, including a two-tab fixture that exercises bridge handshakes programmatically.
  • Converted manual ecosystem reviews into a self-service conformance tool, maintaining a React/NestJS harness over the Allure TestOps API so multiple wallet vendors auto-verify their own SDK integrations.
  • Tamed a ~9-axis combinatorial explosion down to 100+ stable E2E cases across AppKit and Demo Wallet, via shift-left analysis and pairwise design inside a 650+ scenario TestOps repository.

QA Lead — TON Foundation

Apr 2024 – Jan 2026 · Remote · Telegram Apps Center & wider TON ecosystem

  • Validated sub-300ms responses at 50k RPS with no failures observed in peak-load testing, architecting a k6 suite across core flows and partnering with SREs on backend bottlenecks.
  • Cut release delays ~20% by scaling the QA function with metric dashboards and quarterly planning, and authored the QA guidelines that gate ecosystem grants.
  • Saved ~$12k/year in false-positive payouts managing the Hackenproof bug bounty, and co-founded the TON Testers crowd-QA platform to surface regional edge cases.
  • Diagnosed a top Mini App memory leak that was taking down the native Telegram client, and pioneered SSL-pinning bypass via custom Frida scripts for deep network inspection.
  • Automated ecosystem catalog conformance checks with Telethon and Lighthouse, and delivered integration-standard keynotes to 100+ ecosystem developers.

Senior QA Engineer — Whoosh

Dec 2023 – Apr 2024 · High-load micro-mobility · LATAM expansion

  • Validated remote locks and violation detection on real hardware, orchestrating Hardware-in-the-Loop testing with live e-scooters, embedded SIM telemetry, hardware-level GPS spoofing and dynamic route mockers.
  • Safeguarded correctness and reliability of financial transactions on the core billing engine — dynamic fund holds, subscription lifecycles, automated refunds and one-time authorizations.
  • Architected the test frameworks for the LATAM market launch, standardizing QA documentation and mentoring engineers to accelerate regression cycles.

Senior QA Engineer — VK.com

Apr 2021 – Apr 2023 · Remote · Superapp hub & Mini Apps platform · 43M+ MAU, 76k+ apps

  • Shipped the releases I owned on a platform serving 43M+ MAU without high-severity production incidents by moving QA into PRD-stage risk assessment — including blocking unsafe contract updates that would have crashed legacy clients.
  • Owned cross-platform E2E at the frontend-to-network junction (iOS, Android, Web), validating API versioning, backward compatibility and custom REST mechanisms.
  • Drove the mobile release lifecycle via TeamCity build and autotest pipelines with Firebase Crashlytics health monitoring, and authored the Developer Responsibility Matrix that shortened dev-to-QA handoffs.

QA Engineer — CIAN Group

Aug 2019 – Apr 2021 · Moscow · PropTech marketplace, Secondary Real Estate

  • Built the mobile automation framework in Python and Appium with a Page Object architecture, wired into Jenkins CI and unified reporting in Allure TestOps.
  • Owned the platform's primary content path — property listing creation — isolating memory leaks and crashes under massive dynamic forms and heavy media uploads.
  • Led QA for an internal rental-market startup, validating electronic rent execution, document verification and secure transactions; authored API coverage maps and monitored backends via ELK.

Earlier — Mamsy (QA Engineer, 2019) · Redmadrobot (QA Engineer, 2017–2019)

Moscow · E-commerce flash sales · InsurTech at a top-tier product agency · plus QA→automation mentoring, 2021–2023

  • Delivered mobile automation from zero to MVP in Java and Appium at Mamsy, and hands-on-validated Android WMS applications on industrial scanning terminals at the fulfilment centre.
  • Field-tested an InsurTech accident-registration flow under real network degradation at Redmadrobot, isolating vendor-specific defects with ADB and Logcat and running on-site enterprise UAT.

Independent & fractional ventures

Founding QA Architect & SDET Lead — Portex

Apr 2026 – Present · Fractional engagement · Real-money prediction market on Polymarket liquidity

  • Strengthened financial integrity across 6 chains (TON, TRON, EVM, Solana and others) by building a Go API suite in GitLab CI with nightly real-money buy/sell chains and automated on-chain-versus-ledger reconciliation.
  • Closed the parallel over-spend and double-claim class of vulnerability by authoring concurrency tests against the CLOB settlement path, backed by a CI security pipeline on govulncheck, trivy and gosec.
  • Pioneered Agentic QA in production CI — a dual-layer Playwright suite (mock-driven plus live Telegram Web client) with Claude Code prototyping self-repairing tests that adapt to UI drift instead of failing the build.
  • Shipped frontend fixes to production personally and led the engineering team's Jira-to-Linear migration, driving a shift-left culture from inside the Go and React codebases.

Founding QA Engineer & Lead — Portals

May 2025 – Aug 2026 · Independent engagement · #1 Telegram collectibles marketplace · 4.8M users

  • Secured an ecosystem that processed $280M+ in its first year as the 6-person startup's founding QA engineer, establishing the entire quality and security foundation around a unified balance layer.
  • Closed 30+ critical vulnerabilities pre-launch — race conditions, IDOR, RBAC flaws, API data leaks and business-logic abuse — acting as de-facto security engineer. Maintained a record of no known critical vulnerabilities reaching production within my area of responsibility.
  • Proved launch-day capacity on-chain and off by building load infrastructure over Highload V3R1 contracts plus custom Go MTProto scripts that simulate mass concurrent accounts inside Telegram itself.
  • Re-architected the admin/moderator authorization flow and integrated Portals Games, a PvP/PvE ecosystem sharing the marketplace's balance layer.

Co-Founder & Product Owner — TON Testers

Nov 2024 – Dec 2025 · Crowd-QA hub via TON Foundation · In partnership with xcrwd

  • Scaled a distributed testing hub from zero to ~500 vetted crowd-testers, owning product scope, roadmap and backlog for decentralized QA across the TON ecosystem.
  • Wired the tester pool directly into release toggles, catching device fragmentation and localization defects that internal teams could not reproduce.
  • Surfaced production risks at the network layer — Cloudflare caching anomalies and regional geo-blocking that were silently breaking Mini App loading in specific countries.

Public speaking, community & publications

  • Keynote speaker, TON Hackers League — SE Asia tour (2024): two bootcamps, Manila and Ho Chi Minh City, inside a 19-bootcamp, $2M-prize-pool global initiative. Delivered "Telegram Mini Apps — Overview and Best Practices", plus 1:1 mentorship and live wallet-flow debugging with 50+ founders and engineers.
  • Hackathon judge, TON Hackers League (2024): evaluated 50+ startup pitches for a local $5,000 prize pool and a fast-track to the global league, auditing architectural resilience, on-chain settlement correctness, Telegram-native UX and threat models.
  • "Process-as-Code: Managing AI in Web3 SDK Testing" — long-form engineering article, InfoQ draft. Covers control-plane architecture via MCP integrations, why AI should accelerate phase transitions but never authorize them, and a gasless-transfer case study isolating deterministic E2E paths in CI.