Артем

Artem Vashchenko
PROMPT ENGINEER · AI SPECIALIST · QA & TEST BACKGROUND

[відкрити контакти](див. вище в блоці «контактна інформація») [відкрити контакти](див. вище в блоці «контактна інформація») Uman, Ukraine · Remote only LinkedIn / Artem Vashchenko

PROFESSIONAL SUMMARY

IT professional with nearly 5 years of hands-on experience in mobile software and network testing
(ThunderSoft, field engineering), making a deliberate transition into Prompt Engineering and AI automation.
Completed the Prompt Engineer course at GoIT (2025) and applied the knowledge across a series of personal
non-commercial projects.
A QA background is a direct advantage in Prompt Engineering: the ability to write precise test cases translates into
accurate task descriptions for LLMs; experience with log analysis (QXDM, QCAT, ELT) builds the systematic thinking
needed to debug complex prompt chains; structured bug reporting skills provide a disciplined approach to evaluating
expected vs. actual AI model output.

KEY SKILLS

Prompt Engineering AI Tools & LLMs
System Prompt Design · Few-Shot · Zero-Shot GPT-4o · Claude 3 · Gemini
Role Prompting · Tree of Thought · ReAct Midjourney · DALL-E 3 · Notion AI
Prompt Chaining · Negative Prompting LangChain (basics)
Prompt Library Management

Automation & Integrations IT / QA Background
Make · Zapier · n8n Test Cases · Bug Reports · Requirements Analysis
REST API · Webhooks · JSON · Markdown BRD/SRS Documentation · UML · SQL
Basic Python (API requests) QXDM · QCAT · ELT · 2G/3G/4G/5G

WORK EXPERIENCE

Prompt Engineer — Personal Non-Commercial Projects
Self-directed Practice | 2024 – 2025 | Remote
Alongside GoIT coursework, built a series of personal projects to practise real-world LLM application and workflow automation
scenarios.
Project 1: AI Assistant for Job Description Analysis GPT-4o · Python · JSON
–​ Built a prompt system to automatically parse job descriptions: extracted requirements, tech stack, and soft skills into a
structured JSON format
–​ Designed a chain-of-thought prompt: 'what is the company looking for' → 'what is the candidate missing' → final
recommendation
–​ Added few-shot examples for 5 role types (QA, Dev, PM, Data, DevOps) — improved classification accuracy by
~40% vs. zero-shot baseline
–​ Outcome: reduced time to analyse a single job posting from 15 minutes to under 2 minutes

Project 2: QA Documentation Generator Claude 3 · Notion API · n8n
–​ Automated workflow: feature description in → Claude 3 generates test cases in a standardised format → auto-saved
to Notion
–​ Engineered a system prompt with strict output rules (title, reproduction steps, expected result, priority) to meet QA
standards out of the box
–​ Applied role prompting: model acts as a senior QA engineer that asks clarifying questions before generating test
cases
–​ 85% of generated test cases required no manual editing, validated against 30+ real feature descriptions from
ThunderSoft experience

Project 3: Prompt Testing Framework GPT API · Python · Google Sheets
 –​ Applied a QA mindset to prompt evaluation: scoring table across 5 metrics (accuracy, completeness, format,
reproducibility, edge case handling)
–​ Wrote a Python script for batch testing: each prompt runs against 20 input variations; results logged to Google Sheets
for analysis
–​ Identified and documented 12 common prompt anti-patterns — used as a quality standard across all subsequent
projects

Project 4: Image Generation Prompt Library Midjourney · DALL-E 3 · Notion
–​ Structured library of 60+ prompts for branding tasks: logos, banners, illustrations, UI elements, social media visuals
–​ Designed a reusable template with fixed style parameters and variable fields to ensure brand consistency across
outputs
–​ Defined negative prompting rules to eliminate common Midjourney v6 artefacts (extra limbs, blurred text, inconsistent
shadows)

Test Team Leader
ThunderSoft | Aug 2022 – Jun 2025 (2 yrs 11 mos) | Kyiv · Remote
–​ Led a team of testers: task distribution, daily stand-ups, quality control, and blocker resolution
–​ Structured daily progress reporting in a standardised format — a skill directly transferable to documenting AI
workflows and prompt libraries
–​ Log analysis using ELT, QXDM, and QCAT: device activity monitoring, log parsing, and deep diagnostics of complex
technical issues
–​ Detailed bug documentation: reproduction steps, expected vs. actual results, registration and tracking in the bug
tracking system
–​ Cross-team collaboration with development: regular communication on findings, sprint planning participation, and
improvement proposals
–​ Ensured test reproducibility in line with industry standards — a consistency principle carried into prompt engineering
practice

Field Test Engineer
Mobile Industry | Sep 2020 – Aug 2022 (2 yrs) | Kyiv
–​ Mobile network testing (2G/3G/4G/5G): signal strength, connectivity, and performance across diverse environments
and scenarios
–​ VoLTE / VoWiFi / transition tests (SRVCC, CSFB): validation of voice calls, SMS, MMS, and data packages under
various network conditions
–​ Prototype device testing (MTK & Qualcomm chipsets): execution of predefined test cases, software-hardware
compatibility verification
–​ Log capture and analysis (QXDM, QCAT, ELT): device activity recording, log parsing, and targeted data extraction for
issue diagnosis
–​ Mobility and roaming testing: handovers between network cells and regions

EDUCATION & CERTIFICATIONS

Prompt Engineer
GoIT · Online Course · 2025
–​ LLM fundamentals: transformer architecture, tokens, temperature, context window
–​ Prompting techniques: few-shot, chain-of-thought, role prompting, ReAct, tree of thought
–​ Image generation: Midjourney, DALL-E 3, Stable Diffusion — style transfer, negative prompting
–​ AI workflow automation: Make, Zapier, n8n, REST API
–​ Capstone project: AI assistant built for a real business use case

LANGUAGES

Ukrainian English
Native B1/B2 — technical documentation, business correspondence