Сервіс пошуку роботи №1 в Україні
Vitalii
Automation engineer
- Вік:
- 28 років
- Місто проживання:
- Київ
- Готовий працювати:
- Дистанційно
Контактна інформація
Шукач вказав: ТелефонЕл. поштуМесенджер
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/20331601/
Завантажений файл
Версія для швидкого
перегляду
Це резюме розміщено у вигляді файлу. Ця версія для швидкого перегляду може бути гіршою за оригінал резюме.
Vitalii Maslii
Automation & AI Systems Engineer
[відкрити контакти ](див. вище в блоці «контактна інформація») / t.me/vitaliymasliy / [відкрити контакти ](див. вище в блоці «контактна інформація»)
github.com/Monsteroo / vitaliimaslii.com / Ukraine, EET (UTC+2) / remote only
SUMMARY
Self-directed automation and AI engineer. I take systems from architecture to live deployment, with monitoring, health checks and self-
healing restart included. I work on a commercial basis: a live GoHighLevel CRM automation suite for a paying client, and a trading system
running real capital on Binance. Every system below is deployed and monitored, not a sandbox demo — each has real uptime, real
failure modes, and real fixes behind it. Daily hands-on work with Claude Code on production code is my strongest skill and the foundation of
every project here.
SKILLS
AI-native development Claude Code (daily driver, production code every day), Antigravity (hands-on), Cursor (familiar), rapid test-driven AI-native
builds & critical review of AI-generated code
AI & LLM integration Anthropic Claude, OpenAI GPT + Realtime, Google Gemini, Alibaba Qwen — all in production. Prompt engineering, multi-model
orchestration, cost optimisation, vision pipelines, AI agents
RAG & retrieval Hybrid semantic + keyword search, embeddings, vector databases (pgvector, Pinecone), reciprocal rank fusion, confidence gating,
citation validation, reproducible eval harnesses
Automation n8n at expert level (custom nodes, SDK, MCP integrations, multi-branch pipelines with human approval), GoHighLevel CRM (REST + MCP),
webhooks
Backend Python (FastAPI, Telethon, Binance API), PostgreSQL, pgvector, Pinecone, SQLite, Docker & Compose, REST API, WebSocket, pytest, Git
Infrastructure Self-hosted deployment, launchd process management, Tailscale, health-check endpoints, watchdog monitoring and alerting, log-
based diagnostics
Languages Ukrainian (native), English B1 — strong written and reading, async-first
EXPERIENCE
AI Automation & RevOps Engineer — Client Work / Contract 2026 – Present
Delivered a GoHighLevel CRM automation suite for a paying service-business client, eliminating manual follow-ups to 0.
Caught a client-facing logic defect mid-build where customers were promised a callback 30 minutes before staff were told about it.
AI Systems & Backend Engineer — Production Systems & R&D 2025 – Present
Build and operate production Python services (FastAPI, Docker, launchd, Tailscale) with health-check endpoints, watchdog alerting and self-healing
restart — not sandbox demos.
Shipped two production RAG systems with measured retrieval accuracy, 18/18 and 19/20, on reproducible eval harnesses rather than claimed
numbers.
Cut vision-AI inference cost by 70% with no change to output quality, by routing a cheap model for triage and an expensive one only where it
earned its cost.
Military Operations & Logistics Lead — Armed Forces of Ukraine 2022 – 2026
Led operational logistics and resource distribution under strict operational security and high-stress environments.
Accountable for continuous material readiness, mission-critical supply coordination, and structured execution across teams.
SYSTEMS IN PRODUCTION
P3 Trade Executor github.com/Monsteroo/p3-trade-executor 1,249 signals backtested
Binance bot trading real capital. Trend-following strategy validated on 1,249 historical signals before any execution code was written, with a watchdog
that alerts within 10 minutes if the executor goes quiet.
Python · Binance API · Telethon · launchd · n8n
GHL Docs RAG github.com/Monsteroo/ghl-docs-rag 19/20 retrieval accuracy (95%)
Hybrid RAG over 599 chunks of real GoHighLevel documentation. Pinecone dense + sparse vectors in one query, a confidence gate that answers “not
sure” below threshold, and citations verified by code rather than trusted from the model.
Python · FastAPI · Pinecone · BM25 · OpenAI embeddings · Claude
Portfolio RAG Chat vitaliimaslii.com 18/18 retrieval accuracy (100%)
Self-hosted RAG over my own case studies, live on the site. Hybrid retrieval merges pgvector cosine similarity with PostgreSQL full-text search via
reciprocal rank fusion; answers are always cited, and it declines when retrieval is weak.
Python · FastAPI · PostgreSQL · pgvector · Claude · Docker
“Tamara” Real-Estate Monitor github.com/Monsteroo/house-search-bot −70% vision AI cost, same output
Rental monitor over two sources with deduplication, real walking-distance filtering and perceptual-hash fraud detection, on a self-hosted geo stack. A
cheap model triages every photo; a stronger one only sees what earns a second opinion.
Python · FastAPI · PostgreSQL · Docker · Gemini · Claude · OSM · 22/22 tests
Voice Translation for Zoom github.com/Monsteroo/gpt-realtime-translator ~3s round-trip latency
Bidirectional real-time voice translation across three realtime AI APIs in parallel. Most of the engineering was audio, not AI: virtual routing, jitter buffers,
and a choppy-playback bug that traced to Bluetooth switching codecs mid-call.
WebSocket · PyAudio · Qwen · OpenAI Realtime · Gemini Live
CRM Automation Suite GoHighLevel · paying client 0 manual follow-ups after rollout
Production workflows for a service business: missed-call text-back that sends exactly one message to the customer while every escalation stays
internal, plus lead nurture, booking reminders and webhook-based lead capture.
GoHighLevel · Webhooks · n8n · REST API · MCP
Content Factory n8n · human-in-the-loop 34 nodes in one stateless workflow
A topic dropped into Telegram becomes a published LinkedIn post: Claude drafts bilingual copy, Gemini makes the image, a human approves with one
tap. Stateless by design, since n8n's task-runner mode will not persist workflow state.
n8n · Claude API · Gemini · LinkedIn API · Telegram Bot API
Multi-Source Aggregation Bots Python · scheduled 5+ sources in one deduplicated feed
Scrapers across X, Reddit, Telegram and niche job boards, with Claude scoring relevance and scheduling that survives a sleeping laptop. Blocked
datacentre IPs and undocumented APIs meant each source needed its own approach.
Python · Selenium · Claude · SQLite · launchd · n8n
EDUCATION
Kyiv National University of Technologies and Design · BSc & MSc, Computer Science · 2015 – 2020
Full case studies, live systems and the numbers behind them vitaliimaslii.com
Automation & AI Systems Engineer
[
github.com/Monsteroo / vitaliimaslii.com / Ukraine, EET (UTC+2) / remote only
SUMMARY
Self-directed automation and AI engineer. I take systems from architecture to live deployment, with monitoring, health checks and self-
healing restart included. I work on a commercial basis: a live GoHighLevel CRM automation suite for a paying client, and a trading system
running real capital on Binance. Every system below is deployed and monitored, not a sandbox demo — each has real uptime, real
failure modes, and real fixes behind it. Daily hands-on work with Claude Code on production code is my strongest skill and the foundation of
every project here.
SKILLS
AI-native development Claude Code (daily driver, production code every day), Antigravity (hands-on), Cursor (familiar), rapid test-driven AI-native
builds & critical review of AI-generated code
AI & LLM integration Anthropic Claude, OpenAI GPT + Realtime, Google Gemini, Alibaba Qwen — all in production. Prompt engineering, multi-model
orchestration, cost optimisation, vision pipelines, AI agents
RAG & retrieval Hybrid semantic + keyword search, embeddings, vector databases (pgvector, Pinecone), reciprocal rank fusion, confidence gating,
citation validation, reproducible eval harnesses
Automation n8n at expert level (custom nodes, SDK, MCP integrations, multi-branch pipelines with human approval), GoHighLevel CRM (REST + MCP),
webhooks
Backend Python (FastAPI, Telethon, Binance API), PostgreSQL, pgvector, Pinecone, SQLite, Docker & Compose, REST API, WebSocket, pytest, Git
Infrastructure Self-hosted deployment, launchd process management, Tailscale, health-check endpoints, watchdog monitoring and alerting, log-
based diagnostics
Languages Ukrainian (native), English B1 — strong written and reading, async-first
EXPERIENCE
AI Automation & RevOps Engineer — Client Work / Contract 2026 – Present
Delivered a GoHighLevel CRM automation suite for a paying service-business client, eliminating manual follow-ups to 0.
Caught a client-facing logic defect mid-build where customers were promised a callback 30 minutes before staff were told about it.
AI Systems & Backend Engineer — Production Systems & R&D 2025 – Present
Build and operate production Python services (FastAPI, Docker, launchd, Tailscale) with health-check endpoints, watchdog alerting and self-healing
restart — not sandbox demos.
Shipped two production RAG systems with measured retrieval accuracy, 18/18 and 19/20, on reproducible eval harnesses rather than claimed
numbers.
Cut vision-AI inference cost by 70% with no change to output quality, by routing a cheap model for triage and an expensive one only where it
earned its cost.
Military Operations & Logistics Lead — Armed Forces of Ukraine 2022 – 2026
Led operational logistics and resource distribution under strict operational security and high-stress environments.
Accountable for continuous material readiness, mission-critical supply coordination, and structured execution across teams.
SYSTEMS IN PRODUCTION
P3 Trade Executor github.com/Monsteroo/p3-trade-executor 1,249 signals backtested
Binance bot trading real capital. Trend-following strategy validated on 1,249 historical signals before any execution code was written, with a watchdog
that alerts within 10 minutes if the executor goes quiet.
Python · Binance API · Telethon · launchd · n8n
GHL Docs RAG github.com/Monsteroo/ghl-docs-rag 19/20 retrieval accuracy (95%)
Hybrid RAG over 599 chunks of real GoHighLevel documentation. Pinecone dense + sparse vectors in one query, a confidence gate that answers “not
sure” below threshold, and citations verified by code rather than trusted from the model.
Python · FastAPI · Pinecone · BM25 · OpenAI embeddings · Claude
Portfolio RAG Chat vitaliimaslii.com 18/18 retrieval accuracy (100%)
Self-hosted RAG over my own case studies, live on the site. Hybrid retrieval merges pgvector cosine similarity with PostgreSQL full-text search via
reciprocal rank fusion; answers are always cited, and it declines when retrieval is weak.
Python · FastAPI · PostgreSQL · pgvector · Claude · Docker
“Tamara” Real-Estate Monitor github.com/Monsteroo/house-search-bot −70% vision AI cost, same output
Rental monitor over two sources with deduplication, real walking-distance filtering and perceptual-hash fraud detection, on a self-hosted geo stack. A
cheap model triages every photo; a stronger one only sees what earns a second opinion.
Python · FastAPI · PostgreSQL · Docker · Gemini · Claude · OSM · 22/22 tests
Voice Translation for Zoom github.com/Monsteroo/gpt-realtime-translator ~3s round-trip latency
Bidirectional real-time voice translation across three realtime AI APIs in parallel. Most of the engineering was audio, not AI: virtual routing, jitter buffers,
and a choppy-playback bug that traced to Bluetooth switching codecs mid-call.
WebSocket · PyAudio · Qwen · OpenAI Realtime · Gemini Live
CRM Automation Suite GoHighLevel · paying client 0 manual follow-ups after rollout
Production workflows for a service business: missed-call text-back that sends exactly one message to the customer while every escalation stays
internal, plus lead nurture, booking reminders and webhook-based lead capture.
GoHighLevel · Webhooks · n8n · REST API · MCP
Content Factory n8n · human-in-the-loop 34 nodes in one stateless workflow
A topic dropped into Telegram becomes a published LinkedIn post: Claude drafts bilingual copy, Gemini makes the image, a human approves with one
tap. Stateless by design, since n8n's task-runner mode will not persist workflow state.
n8n · Claude API · Gemini · LinkedIn API · Telegram Bot API
Multi-Source Aggregation Bots Python · scheduled 5+ sources in one deduplicated feed
Scrapers across X, Reddit, Telegram and niche job boards, with Claude scoring relevance and scheduling that survives a sleeping laptop. Blocked
datacentre IPs and undocumented APIs meant each source needed its own approach.
Python · Selenium · Claude · SQLite · launchd · n8n
EDUCATION
Kyiv National University of Technologies and Design · BSc & MSc, Computer Science · 2015 – 2020
Full case studies, live systems and the numbers behind them vitaliimaslii.com
Схожі кандидати
-
Тестувальник
45000 грн, Дистанційно -
QC engineer
Дистанційно -
Тестувальник
Дистанційно -
Strong Middle QA Engineer (Automation, Manual)
45000 грн, Дистанційно -
Тестувальник
20000 грн, Дистанційно -
QA-інженер
Дистанційно, Чернігів