Contact information

You can get this candidate's contact information from https://www.work.ua/resumes/20709583/

Work experience

Prompt Engineer / AI Automation Specialist

from 02.2025 to now (1 year 9 months)
Self-employed / private clients, Дистанційно (Приватні особи)

- Translate loosely defined business requirements into structured AI workflows, acceptance criteria and working prototypes.
- Build document-processing pipelines with 10-50+ sequential LLM steps, covering source analysis, data normalization, content generation and final validation.
- Connect external services through REST APIs and webhooks; work with n8n, OpenAI API, Claude, Ollama, MCP, Google Sheets/Drive and email automation.
- Use Codex, Claude and other AI assistants to research solutions, read documentation, analyze code, develop workflow scenarios and check edge cases.
- Develop Python scripts and browser automation with Playwright; troubleshoot workflows using logs, DOM inspection and technical documentation.
- Design evaluation sets, regression checks, quality checklists and fallback logic to improve the consistency and repeatability of AI outputs.
- Work with RAG, embeddings and chunking to organize source material and supply relevant context to language models.
- Build and maintain public projects, including an AI client-intake CRM, an offline Markdown-to-Word Chrome extension, a reusable Moodle automation instruction package and a structured prompt library.
- Consider privacy and prompt-injection risks when designing AI-assisted workflows.

Education

National University of Food Technologies

Bachelor's degree in Biotechnology and Bioengineering (specialty 162), Київ
Higher, from 2022 to 2026 (4 years)

Faculty of Biotechnology and Environmental Control. Bachelor's degree completed in 2026.

Additional education and certificates

Udemy - Automate Everything: n8n, LLMs, OpenAI API, Claude, Ollama, MCP & RAG

2025

Google / Ministry of Economy of Ukraine / EdEra - AI for Business

2025

Ukraine Global Faculty / KIEF Academy - Autonomous AI Agents: Use Cases

2025, seminar

Knowledge and skills

• Rag
• Reverse engineering
• Python Requests
• LLM
• N8n
• Ollama
• API integration
• Webhooks
• Harness engineering
• AI prompting
• AI integration
• LLM agents
• Python
• OpenAI API
• AI automation
• Prompt engineering
• LLM evaluation
• Context engineering
• Codex
• Claude
• MCP
• REST API
• Playwright
• Browser automation

Language proficiencies

• English — above average
• Ukrainian — fluent

Additional information

PROFESSIONAL SUMMARY
AI Automation Specialist and Prompt Engineer with hands-on experience since February 2025 in LLM workflows, API integrations and document-processing automation. Turn loosely defined requirements into structured workflows and working prototypes. Use Python, n8n, OpenAI API, Claude, Codex and external APIs to automate routine tasks, investigate issues and validate results. Focus on repeatable execution, clear acceptance criteria and practical business value. Seeking an AI Engineer role building useful AI integrations and business-process automation.

TECHNICAL PROFILE
AI and agents: Claude, Codex, OpenAI API, Ollama, MCP, prompt engineering, context engineering and multi-step prompt chains.
Automation and integration: n8n, REST APIs, webhooks, Google Sheets/Drive, email workflows and browser automation.
Development: Python, Python Requests, Playwright, JavaScript basics, HTML/CSS, DOM inspection and debugging with logs and documentation.
CRM portfolio stack: Next.js, React, TypeScript and Supabase Postgres/Auth, with row-level security and transactional RPCs.
Knowledge and quality: RAG, embeddings, chunking, evaluation sets, regression checks, acceptance criteria, fallback logic and awareness of privacy/prompt-injection risks.

SELECTED PUBLIC PROJECTS

Intake - AI client-intake CRM and business-process automation
https://github.com/kokojas/client-intake-crm
Built a portfolio MVP that saves client inquiries, creates contacts and deals, and uses the OpenAI Responses API with Structured Outputs to extract a reviewable brief, requirements, budget/deadline and follow-up tasks. Implemented duplicate-event protection, an activity history, owner-only Supabase access and server-side AI spending controls. Preserved original messages on AI failures and cached completed analyses to avoid repeated paid calls. Verified the workflow with 24 automated tests, database checks and browser testing. Stack: Next.js, React, TypeScript, OpenAI API and Supabase.

MD to DOCX (Offline) - Chrome extension and document-processing product
https://chromewebstore.google.com/detail/md-to-docx-offline/fcidkfjmabiklghkjfionfchkpidaefi
Developed a Manifest V3 Chrome extension that converts Markdown to Word locally, including LaTeX-to-OMML equations and PDF text/OCR import. Separated the offline converter from optional online AI tools backed by a Python/Flask API, with provider keys and usage quotas handled server-side. Added selected-text rewriting and regression tests for conversion and editor behavior, and took the product through to a usable public release.

NUFT Moodle Automation - AI agent instruction package and browser workflow
https://github.com/kokojas/nuft-moodle-dom-copy
Created a reusable instruction package for AI agents to run a full Moodle workflow: scrape page content and assignment requirements, analyze the source material, decompose the task into stages, execute and validate each stage, and automatically upload completed files to the platform. Used Codex/Claude Code and Playwright for browser interaction. Published a sanitized DOM copy with representative course and assignment-upload flows for development and testing, without credentials or personal data.

Prompt Library - reusable prompt collection
https://github.com/kokojas/prompt-library
Organized 54 reusable prompts (34 in English and 20 in Ukrainian) by purpose, required inputs and expected outputs to support reuse and evaluation.

WORKING APPROACH
Break complex tasks into verifiable steps, check assumptions and edge cases, and use evaluation/regression checks rather than relying on a single successful prompt. Comfortable learning from technical documentation and iterating on prototypes in response to feedback.

LANGUAGES
English - Upper-Intermediate; Ukrainian - fluent.

AVAILABILITY
Full-time remote work.