Сервіс пошуку роботи №1 в Україні
Максим
AI developer
- Вік:
- 21 рік
- Місто проживання:
- Ужгород
- Готовий працювати:
- Дистанційно
Контактна інформація
Шукач вказав: ТелефонЕл. пошту
Прізвище, контакти та світлина доступні тільки для зареєстрованих роботодавців. Щоб отримати доступ до особистих даних кандидатів, увійдіть як роботодавець або зареєструйтеся.
Отримати контакти цього кандидата можна на сторінці https://www.work.ua/resumes/19763707/
Завантажений файл
Версія для швидкого
перегляду
Це резюме розміщено у вигляді файлу. Ця версія для швидкого перегляду може бути гіршою за оригінал резюме.
POLUSMIAK MAKSYM
[відкрити контакти ](див. вище в блоці «контактна інформація») • [відкрити контакти ](див. вище в блоці «контактна інформація») • LinkedIn • GitHub • Uzhhorod, Ukraine
SUMMARY
AI developer building LLM agents and automation: a multi-agent lead pipeline with an evaluation gate, a voice-agent platform on
LiveKit, and an end-to-end job-application automation system run by AI agents.
SKILLS
AI / ML: LLM APIs, AI Agents, Prompt Engineering
DevOps / Tools: n8n, Docker, Git, GitHub, Docker Compose
Backend: Node.js, FastAPI, REST API, API Integration
AI Providers: OpenAI, Claude, Google Gemini
Programming: Python, TypeScript, JavaScript, HTML, CSS
Databases: SQLite, PostgreSQL, Supabase, MongoDB
AI Frameworks: MCP, LangChain
Frontend: Next.js, Tailwind CSS
Other: OpenRouter API
Human Languages: Ukrainian (native), English (B2)
PROJECTS
Fleet — Multi-Agent Dev Orchestrator 2026
Technologies: AI Agents, Prompt Engineering, LLM APIs, Python, Claude, OpenAI, Google Gemini, Git
Built and open-sourced Fleet, a Claude Code / Codex plugin that plans, implements, reviews and releases Git tasks with multiple
AI workers: six agent skills, a deterministic Python helper and commit-linked check evidence with secret redaction; installs on
Claude Code, Codex and Antigravity.
Built an MVP of a multi-agent development orchestrator: control-plane API with task leases and attempts, parallel coding agents
across Claude, Codex and Gemini, cross-model code review, quota-aware scheduling and a Telegram/web operator panel.
LeadFlow — Multi-Agent AI Lead Pipeline 2025 – 2026
Technologies: LLM APIs, n8n, Node.js, TypeScript, Shopify GraphQL, Telegram Bot API, Langfuse, Ajv
Built a multi-agent AI lead-processing system in n8n: Gmail/webhook intake, intent and language classification, Shopify
enrichment and automated response drafting.
Developed and packaged a custom TypeScript n8n node (n8n-nodes-llm-guard) plus an offline evaluation harness for
reproducible AI workflow behaviour.
Made the pipeline lose nothing silently: every AI answer is validated against a JSON Schema and repaired or rejected, failed
leads go to a dead-letter queue with a Telegram alert and one-command replay; verified with 6 deliberate failure scenarios.
Voxora — AI Twin Platform 2025 – 2026
Technologies: Python, FastAPI, Next.js, TypeScript, Supabase/PostgreSQL, OpenRouter, OAuth2
Integrated Gmail via OAuth2 so agents read and act on real mailbox data with structured LLM outputs.
Voice Fleet — Voice-Agent Platform 2025 – 2026
Technologies: Python, FastAPI, PostgreSQL, LiveKit, Deepgram, Gemini, Cartesia, Docker
Built Voice Fleet, a real-time voice AI agent platform (LiveKit, Deepgram, Gemini, Cartesia) with versioned agent configs and an
operator console (transcripts, events, diagnostics).
Delivered live voice conversations with the agent through LiveKit, wiring speech-to-text, LLM and text-to-speech providers end to
end.
EDUCATION
Uzhhorod National University 2022 – 2026
Computer Engineering
CERTIFICATES
EF SET English Certificate (B2)
Claude Code in Action
Introduction to Model Context Protocol
Model Context Protocol: Advanced Topics
[
SUMMARY
AI developer building LLM agents and automation: a multi-agent lead pipeline with an evaluation gate, a voice-agent platform on
LiveKit, and an end-to-end job-application automation system run by AI agents.
SKILLS
AI / ML: LLM APIs, AI Agents, Prompt Engineering
DevOps / Tools: n8n, Docker, Git, GitHub, Docker Compose
Backend: Node.js, FastAPI, REST API, API Integration
AI Providers: OpenAI, Claude, Google Gemini
Programming: Python, TypeScript, JavaScript, HTML, CSS
Databases: SQLite, PostgreSQL, Supabase, MongoDB
AI Frameworks: MCP, LangChain
Frontend: Next.js, Tailwind CSS
Other: OpenRouter API
Human Languages: Ukrainian (native), English (B2)
PROJECTS
Fleet — Multi-Agent Dev Orchestrator 2026
Technologies: AI Agents, Prompt Engineering, LLM APIs, Python, Claude, OpenAI, Google Gemini, Git
Built and open-sourced Fleet, a Claude Code / Codex plugin that plans, implements, reviews and releases Git tasks with multiple
AI workers: six agent skills, a deterministic Python helper and commit-linked check evidence with secret redaction; installs on
Claude Code, Codex and Antigravity.
Built an MVP of a multi-agent development orchestrator: control-plane API with task leases and attempts, parallel coding agents
across Claude, Codex and Gemini, cross-model code review, quota-aware scheduling and a Telegram/web operator panel.
LeadFlow — Multi-Agent AI Lead Pipeline 2025 – 2026
Technologies: LLM APIs, n8n, Node.js, TypeScript, Shopify GraphQL, Telegram Bot API, Langfuse, Ajv
Built a multi-agent AI lead-processing system in n8n: Gmail/webhook intake, intent and language classification, Shopify
enrichment and automated response drafting.
Developed and packaged a custom TypeScript n8n node (n8n-nodes-llm-guard) plus an offline evaluation harness for
reproducible AI workflow behaviour.
Made the pipeline lose nothing silently: every AI answer is validated against a JSON Schema and repaired or rejected, failed
leads go to a dead-letter queue with a Telegram alert and one-command replay; verified with 6 deliberate failure scenarios.
Voxora — AI Twin Platform 2025 – 2026
Technologies: Python, FastAPI, Next.js, TypeScript, Supabase/PostgreSQL, OpenRouter, OAuth2
Integrated Gmail via OAuth2 so agents read and act on real mailbox data with structured LLM outputs.
Voice Fleet — Voice-Agent Platform 2025 – 2026
Technologies: Python, FastAPI, PostgreSQL, LiveKit, Deepgram, Gemini, Cartesia, Docker
Built Voice Fleet, a real-time voice AI agent platform (LiveKit, Deepgram, Gemini, Cartesia) with versioned agent configs and an
operator console (transcripts, events, diagnostics).
Delivered live voice conversations with the agent through LiveKit, wiring speech-to-text, LLM and text-to-speech providers end to
end.
EDUCATION
Uzhhorod National University 2022 – 2026
Computer Engineering
CERTIFICATES
EF SET English Certificate (B2)
Claude Code in Action
Introduction to Model Context Protocol
Model Context Protocol: Advanced Topics
Схожі кандидати
-
AI-розробник
30000 грн, Дистанційно -
AI-розробник
Дистанційно -
Full stack software engineer
Дистанційно -
AI-розробник
Дистанційно, Ужгород -
AI engineer
Дистанційно, Дніпро, Кам'янське -
Full stack програміст
Дистанційно, Київ , ще 3 міста