Skip to work
DG logo
Dimosthenis Gkontolias
Software engineer + AI builderAthens, Greece

Dimosthenis Gkontolias

I build software that gets past the demo: QR menus in cafes, GCP data pipelines, RAG dashboards, and the content that helps people use them.

EU citizen and Computer Science graduate from AUEB (9.15/10, valedictorian), founder of Quar.gr, former Erasmus+ data intern at TrackSights, and creator behind Demos Vibes. My best work sits where code, product sense, data, and distribution meet.

See the workDownload CV

9.15/10

AUEB GPA

BSc Informatics, valedictorian and top of department.

10+

cafes in production

Quar.gr is live with QR menus, admin editing, and analytics.

785k

training listings

TrackSights depreciation model over 71 features, R² = 0.946.

60+

projects shipped

Web apps, data pipelines, AI tools, mobile apps, and games across 3+ years.

Recruiter snapshot

What a recruiter can learn fast.

A quick read before the project cards.

Best fit

Product-focused engineering roles, especially React/Next, backend/data work, and AI features that need a clean path from prototype to production.

Proof

Quar.gr is used by cafes, TrackSights work touched real cloud data pipelines, and Demos Vibes makes me explain tools in plain Greek.

Why me

I am still early, but I have already dealt with users, messy data, deploys, content, and the uncomfortable parts after a demo works.

Work setup

EU citizen based in Athens; Greek native, English C2/ECPE, and German B2/OSD.

Evidence map

What each project proves.

A faster recruiter read: proof first, then the hiring signal behind it.

Quar.gr

Production ownership

Proof

10+ cafes use the product; admin editing, media, analytics, deploys, and support are already part of the loop.

Recruiter read

Can own a user-facing product after the demo, including the unglamorous operational parts.

View project

TrackSights

Data engineering depth

Proof

OEM providers, Pydantic schemas, GCS, BigQuery, Dataform, Terraform, assertions, and a 785k-listing model dataset.

Recruiter read

Comfortable turning messy external data into warehouse-ready datasets and model features.

View project

DataViz

AI product UX

Proof

RAG over 23,000 data points and 207 metrics with streaming answers and generated chart output.

Recruiter read

Can wrap model behavior in a product surface that makes data exploration understandable.

View project

Demos Vibes

Distribution ability

Proof

Short-form Greek AI demos backed by reusable resource pages, typed content objects, SEO, sitemap, and JSON-LD.

Recruiter read

Can explain technical workflows clearly and turn content feedback into better product assets.

View project

Nero Website

Frontend polish + QA

Proof

Real venue assets, QuaR menu path, structured data, image gallery, mobile interactions, and Playwright checks.

Recruiter read

Can ship a small commercial frontend that looks real, loads assets, and survives basic regressions.

View project

Signal ledger

Quiet proof that the work ethic is not new.

A compact layer for the academic and competition signals that recruiters often scan for after the projects.

Panhellenic exams

19,150/20,000

National exam score in 2021-2022, before entering AUEB Informatics.

Model school finish

19.3/20

Graduated from Anavrita Model Public High School with a 19.3/20 grade.

Hackathon finals

2x finalist

Finalist at Hamburg ChefTreff Hackathon and Athens AceIn AI Hackathon.

Role fit

Where I can be useful fastest.

Three recruiter-readable paths, each backed by work already visible on the page.

Product engineering

Best fit when a team needs someone who can take a product surface from unclear requirements to something users can actually operate.

  • Quar.gr: product, admin UX, data model, deploys, support.
  • Nero Website: polished hospitality frontend with Playwright QA.

AI and data products

Strongest when the work mixes messy data, model behavior, cloud jobs, and a frontend that makes the result understandable.

  • TrackSights: OEM providers, GCP pipeline, BigQuery/Dataform.
  • DataViz: RAG chatbot over 23,000 points and 207 metrics.

Distribution and education

Useful for teams that want engineering plus the ability to explain, publish, and learn from audience feedback.

  • Demos Vibes: Greek AI demos with reusable resource pages.
  • Creator loop: build the tool, explain the workflow, improve the asset.

Operating mode

How I ship when things are messy.

The pattern behind the projects: make the work legible, test what can break, and keep the loop close to real users.

Make the messy path usable

I turn unclear requirements into routes, schemas, admin flows, and QA paths that a real operator can trust.

Quar.gr moved from menu idea to cafe-ready admin editing, media, analytics, and support.

Design data before dashboards

I care about source quality, canonical schemas, assertions, and review labels before the chart layer gets pretty.

TrackSights used provider patterns, Pydantic schemas, BigQuery/Dataform, and model features at scale.

Ship with receipts

I prefer visible demos, public links, tests, structured metadata, and clean diffs over claims that cannot be inspected.

Nero and this portfolio both carry Playwright coverage for SEO, assets, overflow, links, and motion preferences.

Explain what I build

I can translate technical workflows into plain Greek content, which makes tools easier to adopt and improves feedback.

Demos Vibes turns short AI demos into reusable pages, not throwaway posts.

Selected work

Projects that shipped beyond the repo.

A recruiter should be able to see the role, stack, constraint, and result quickly. These are the clearest examples.

Founder / production SaaS2025 - PresentLive in production

Quar.gr

A QR menu platform for cafes with admin editing, media management, multilingual menus, and analytics.

Role: Designed and built the product end-to-end: React/TypeScript frontend, Fastify backend, Firebase data layer, admin flows, deployments, and customer-facing UX.

Quar.gr product homepage showing a QR menu phone preview

Why it matters

  • Serving 10+ cafes in production, with 300+ commits behind the product.
  • Admin panel supports drag-and-drop menu editing, images, multilingual content, and GA4 analytics.
  • Real business constraints: onboarding, reliability, content updates, and support for non-technical owners.
ReactTypeScriptFastifyFirebaseGA4Tailwind
LiveGitHub
Erasmus+ data internshipFeb - Jun 2026Cloud data pipeline

TrackSights OEM data pipeline

Provider-based data collection for official car configurator data and a depreciation model for automotive pricing.

Role: Built brand providers, normalized messy OEM data into a canonical schema, and helped turn raw JSON into warehouse-ready datasets.

Why it matters

  • Covered Audi, Hyundai, and Mercedes-Benz across 8 European markets.
  • Pipeline used Cloud Run Jobs, Docker, GCS, BigQuery, Dataform, Secret Manager, Terraform, and data-quality assertions.
  • Modeled depreciation on roughly 785k listings with 71 features using CatBoost, LightGBM, and Optuna, reaching R² = 0.946 on held-out data.
PythonPydanticGCPBigQueryDataformTerraformCatBoost
Case study available on request
Gen AI economic insight platform2026Live demo

DataViz

A full-stack RAG chatbot for economic data exploration with streaming answers and chart generation.

Role: Built the FastAPI + React/TypeScript application, connected Gemini 2.5 Pro, and designed the data-to-chart interaction model.

DataViz Talk To Greek Data landing screen

Why it matters

  • Works over 23,000 data points and 207 metrics.
  • Streams responses with SSE and can render 12 chart types.
  • Turns static economic datasets into an interactive analysis surface.
FastAPIReactTypeScriptGeminiRAGSSE
Live
Content creator + resource hub2026Active channel

Demos Vibes

A Greek AI tools and workflows channel paired with a static resource hub for every video.

Role: Created the content architecture, the demosvibes.gr site, searchable resource pages, SEO metadata, JSON-LD, and the short-form education loop around it.

Demos Vibes resource hub with Greek AI workflow pages

Why it matters

  • 25 recent Instagram reels and 21 TikTok videos around practical AI tools and workflows.
  • Companion hub is data-driven: one typed resource object powers routes, cards, SEO, sitemap, and JSON-LD.
  • Shows an ability recruiters rarely see in junior engineers: build the product and explain why it matters.
ReactTypeScriptVite SSGSEOJSON-LDContent architecture
HubGitHub
AI/event discovery engine2026Private repo

CSEventFinder

A high-recall tracker for computer science, startup, fintech, AI, developer, conference, and hackathon events in Greece.

Role: Built the crawl, extraction, scoring, review-label, SQLite, FastAPI, and Vite app loop.

Why it matters

  • Extracts candidates from JSON-LD, feeds, ICS links, anchors, and broad discovery queries.
  • Scores relevance, date confidence, location, source quality, and learned label signals.
  • Designed for reviewability: false positives are acceptable if evidence is explainable.
PythonFastAPISQLiteViteTypeScriptCrawling
Case study available on request
Privacy-first browser tool2026Open source

QR Style Studio

A client-side QR style generator with custom shapes, frames, logos, export options, and scannability safeguards.

Role: Built the browser-only generator, export pipeline, logo sanitization, SVG escaping, and security-focused UI.

Why it matters

  • No backend, accounts, tracking, cookies, or runtime external origins.
  • Validates logo uploads with type allow-lists, magic-byte checks, size limits, and canvas re-encoding.
  • Exports PNG, JPEG, and SVG while warning about dangerous payloads.
TypeScriptReactSVGCanvasCSPDOMPurify
GitHub

Project shelf

More shipped experiments.

Smaller or less public, but useful for range: AI review tools, hospitality frontends, game backends, and product experiments.

AI code review CLI2026Open source experiment

BoomAI

AI-powered code-review and auto-fixing CLI for C#/Unity projects using static analysis plus LLM patch planning.

Role: Combined Semgrep, DevSkim, Roslyn-style checks, Gitleaks, chunked review planning, and patch application into one CLI flow.

Why it matters

  • Targets bugs such as resource leaks, async pitfalls, null refs, race conditions, and security issues.
  • Designed around exact code replacements and auditable diffs.
PythonGeminiStatic analysisC#Unity
GitHub
Hospitality frontend2026Public GitHub Pages

Nero Website

A polished cafe/bar/restaurant website for Nero in Neo Irakleio, built around real venue photography, fast menu/map actions, and QuaR-backed menu paths.

Role: Designed and built the React/Vite/TypeScript experience, structured data, image gallery, language picker, motion layer, and Playwright QA around verified venue facts.

Nero Website hero section with real daytime lounge photography

Why it matters

  • Published on GitHub Pages with direct QuaR menu, map, call, hours, and social paths.
  • Uses sourced and refined real Nero assets instead of generated imagery, with image preview and mobile swipe interactions.
  • Protected by Playwright checks for overflow, structured data, favicon, image loading, accessibility labels, and external-link rel hygiene.
ReactTypeScriptVitePlaywrightSEOGitHub Pages
LiveGitHub
Game backend2025Client/team project

AthensSecret VR/MR backend

Backend API for a multiplayer VR/MR game set around Athens secrets.

Role: Worked on authentication and backend services connecting the game experience to persistent data.

Why it matters

  • Built with ASP.NET Core, JWT authentication, PostgreSQL, and Dockerized services.
  • Connected Unity/game work with web/backend infrastructure.
ASP.NET CoreC#JWTPostgreSQLDocker
GitHub

Experience

Where the habits came from.

Across internships, Quar.gr, and freelance work, the pattern is the same: own the unclear part, make it reliable, and keep the product understandable.

Data Analyst Intern

TrackSights ApS | Denmark / Erasmus+
Feb 2026 - Jun 2026

Worked in the data team on automotive OEM collection and depreciation modeling.

  • Reverse-engineered configurator data for Audi, Hyundai, and Mercedes-Benz across BE, DE, DK, ES, FR, IT, NL, and SE.
  • Built provider-pattern scrapers and canonical Pydantic schemas, then moved JSON through GCS, BigQuery, and Dataform.
  • Added warehouse assertions and worked on a depreciation model trained over roughly 785k listings and 71 features, reaching R² = 0.946.

Creator layer

I build the thing, then explain it in Greek.

Demos Vibes is my public lab for AI tools, workflows, and reusable resources. Every short demo points back to a resource page, so the content is not just a clip; it becomes something people can use again.

Visit demosvibes.grWatch reels
Instagram

@demos.vibes

Greek AI tool and workflow demos; 25 recent reels.

TikTok

@demos.vibes

Short-form AI demos paired with reusable resources.

X

@demosvibes

AI tools, build notes, and product experiments.

Threads

@demos.vibes

Creator channel presence for Greek AI education.

Stack

The stack I actually use.

This is not keyword stuffing. It is the stack behind the projects above: product UI, APIs, scraping, ML/data work, cloud jobs, and the pieces needed to ship.

Languages

PythonTypeScriptJavaScriptJavaC#C++SQLRDart

Frontend

ReactNext.jsTailwind CSSshadcn/uiViteFramer Motion

Backend and data

FastAPINode/ExpressDjangoFlaskASP.NET CorePostgreSQLBigQueryRedis

AI and automation

RAGLLM APIsPyTorchscikit-learnPlaywrightSeleniumAI coding agentsCursorGitHub Copilot

Cloud and ops

DockerTerraformGCPFirebaseGitCI/CDDataform

Recruiter FAQ

Quick answers without the scroll hunt.

The questions I would expect someone to ask after scanning the projects, answered directly.

What roles is Dimosthenis Gkontolias strongest for?

Product-focused software engineering, AI/data engineering, and frontend-heavy roles where React, TypeScript, Python, FastAPI, GCP data work, and practical AI features meet real users.

Which projects should a recruiter inspect first?

Start with Quar.gr for production ownership, TrackSights for data engineering depth, DataViz for AI product UX, Demos Vibes for distribution ability, and Nero Website for frontend polish with QA.

Is Dimosthenis also a content creator?

Yes. Demos Vibes is his Greek AI tools and workflows channel, with Instagram, TikTok, X, and Threads handles plus a reusable resource hub at demosvibes.gr.

Where is he based and what languages does he work in?

He is an EU citizen based in Athens. He is a native Greek speaker, has English C2/ECPE, and German B2/OSD.

How can recruiters contact him?

The fastest path is email at dgodolias18@gmail.com or LinkedIn. The portfolio also includes a downloadable CV and a vCard contact file.

Next role

I am looking for a team where shipping matters.

Open to software engineering or AI/data engineering roles where I can ship, learn fast, and stay close to users.

Email meLinkedInSave contact