IGOR GLAR / SYSTEMS DEVELOPER

[CHATHELPER EXT] SYS_OK

Logic
before
pixels

I build systems that solve real problems,
then design interfaces that make them simple.

AI does the heavy lifting where it's proven. Judgment calls stay mine.

VUE / NUXT / TYPESCRIPT / AI

SEE CASE STUDY

Igor Glar Portfolio — Logic Before Pixels 3D Architectural Composition

SCROLL

02 / EVIDENCE

evidence.resolve()

More than execution.
I solve the real problem

Every project starts with questioning the surface request. I identify root operational bottlenecks and engineer AI pipelines, vector search systems, and automated workflows that deliver measurable ROI.

01 pipeline.run()

REQUEST

“Upload product photos from suppliers”

WHAT I SAW

Source images in different quality and formats. Store looks unfinished.

SOLUTION

Built a bulk AI pipeline for image processing and infographic overlay.

RESULT

+24% conversion rate

Saved hundreds of hours of designer work.

02 vector.query()

REQUEST

“Build search across our knowledge base”

WHAT I SAW

Keyword search fails. Clients don't write in formal language.

SOLUTION

Vector RAG search + AI chat in operator panel.

RESULT

3x faster response time

Lower escalation rate.

03 bot.sync()

REQUEST

“Need a Telegram bot for requests”

WHAT I SAW

Manual requests get lost in chats. No status tracking.

SOLUTION

Telegram bot with auto-sync to CRM and notifications.

RESULT

0 lost leads

Transparent sales funnel

03 / FLAGSHIP

products.deploy()

Architected for impact.
Built to scale

01 / PRODUCT radar.init()

KEY MECHANICS & INNOVATIONS

AI Memory Decks: Flux-powered engine transforming photos & stories into dual AI portraits (Pre-Duo Roots, Inside Jokes, Future Visions).
Daily Questions Engine: 1000+ deep psychological questions with blind answers and shared relationship timeline.
End-to-End Encryption: Signal-grade privacy architecture securing personal answers, photos, and messages.

TECH STACK

REACT NATIVE EXPO SUPABASE FLUX API NODE.JS E2E ENCRYPTION

02 / TOOL copilot.start()

WHAT WAS IMPLEMENTED

Chrome Extension + Web Panel: Unified operator workspace built to eliminate cognitive overload.
Vector RAG Search: Instant script lookup matching informal customer wording with formal knowledge base docs.
AI Co-pilot: Built-in assistant for polite response generation, automated hints, and fast junior onboarding.

TECH STACK

CHROME EXTENSION API VUE.JS PYTHON VECTOR SEARCH LLM APIS

04 / APPROACH

process.execute()

How I work.
From problem to production

01

Product thinking, not blind execution

I look past the ticket to the business goal. If there’s a simpler, cheaper, more effective path — I raise it before budget is spent, not after.

02

One pipeline, not four handoffs

Scraping, cleaning, embeddings, RAG, reranking, UI — usually four specialists and three handoffs. I run it as one connected system.

03

MVP speed without waiting on design

I design the UX myself from proven patterns, architect for growth, and cut what isn’t needed to ship fast.

05 / STACK

sys.inspect()

Selected tools for speed & scale

01

FRONTEND

Vue 4.5+ Nuxt 4.5+ TypeScript Tailwind CSS Pinia HTML5 / CSS3

02

BACKEND

Node.js Python Supabase PostgreSQL REST & GraphQL

03

AI & DATA

Vector Search (RAG) LLM Integration Flux API Web Scraping Embeddings

06 / CONTACT

signal.dispatch()

Got something
to build?

Direct line. No middle management, no unnecessary handoffs.

@igorglar ↗

EMAIL

igor@glar.space ↗

Single Grounded 3D Dark Charcoal Cube — Compositional Closure for Igor Glar Contact Section

Logic
before
pixels