Skip to content
Vector
HomeProjectsServicesAboutContact
ENUARU
VectorAutomation & AI integrations

© 2026 Vector Studio. All rights reserved.

Pages

  • Home
  • Projects
  • Services
  • About
  • Contact

Contact

  • vectorstudio.ai1@gmail.com
  • Telegram
  • +380 63 396 60 22

More

  • Upwork profile
  • Privacy

Projects

Real projects

Problem → solution → result → stack. Systems I built and run.

Hover a screen · click to jump to its project

  • Job Radar
  • 126-Page Parts Catalog
  • Marketplace Product Pipeline
  • Job Radar
  • 126-Page Parts Catalog
  • Job Radar
JRJob Radar
  1. 3. [UW] SaaS Full Stack Developer for Next.js, Node.js and AI Improvements

    5 жовт, 22:23 · відгуків 61

    бал 5

  2. 4. [UW] Animated SVG hero for a Next.js homepage: ancient harbor line drawing (motion + front-end)

    $600 fixed · 5 жовт, 21:09 · відгуків 21

    бал 5

  3. 5. [UW] Full-Stack Developer — React / TypeScript / Node.js / Azure

    5 жовт, 18:54 · відгуків 73

    бал 5

  4. 6. [FH] Допомога з магістерським проектом Розробка інтелектуального чат-бота для клієнтських консультацій

    10000 UAH · 6 жовт, 09:45 · відгуків 21

    бал 4

01Automation + AI

Job Radar

Checking several job boards by hand took time, and most listings were off-topic, underpriced or spam. I built scheduled n8n workflows that pull jobs from Freelancehunt and Upwork, score each one and send the best to Telegram twice a day.

  • Role: design & build
  • Stack: n8n, PostgreSQL, Telegram
  • Running since October 2026
daily digests
2×
health checks
15 min
Read the case
  1. Freelancehunt + Upwork
  2. Scoring
  3. Telegram
  4. PostgreSQL
Supplier: HW - drive shaftMarketplace titleКарданный вал для автомобиля Iveco

price = rate × supplier price, rounded up

02E-commerce

Marketplace Product Pipeline

Supplier data wasn't storefront-ready: titles, descriptions and prices needed manual work for every product. I built an n8n pipeline that validates each item, rewrites it with a local LLM and prices it by a fixed formula before publishing.

  • Role: design & build
  • ~200 products
  • Local LLM
products
~200
validate → rewrite → price
3 steps
Read the case
  1. Supplier data
  2. Validation
  3. Local LLM
  4. Marketplace

03Data → Print

126-Page Parts Catalog

A 126-page technical catalog had to be translated and laid out without losing a single part number. I wrote a generator that translates with a curated glossary, lays out the pages and checks every part number against the source.

  • Role: generator & QA
  • 126 pages
  • 1,954 part numbers
pages
126
verified
1,9541,954 / 1,954
critical
0
Read the case
  1. Source data
  2. Glossary DE→UK
  3. QA
  4. PDF

Your project could be the next one.

Start a project

n8n automation from $100 · Telegram bot from $150 · Website review from $80