Europe/Munich Projects / Skills / Evidence

Technical Project Portfolio

Eray Kuyumcu development work, explained through technical evidence.

Development portfolio documenting practical technical work across product, backend, mobile, AI, and operations.

Current focus

Product engineeringBackend systemsMobile appsAI toolingQuality reviews

This is a focused portfolio of technical project case studies, not a complete employment history. The site is structured around evidence: projects, decisions, and skills while avoiding product-control claims and sensitive operational details.

Profile

Evidence-led

Selected technical work is organized so each claim can point to project and skill evidence.

Assistant

Evidence-backed

The ask page checks a visitor project request against public project and skill evidence.

Case Studies

Expandable

Project pages and skill data share the same content system for future search and citation.

Selected work

Project case studies

These are live production applications. Source code is private, so the case studies focus on stack, decisions, implementation patterns, and outcomes.

A cross-platform photo system case study spanning a Bun/Hono API, Next.js web interface, SwiftUI iOS app, Kotlin Android app, realtime infrastructure, AI photo analysis, moderation workflows, private messaging, and production-readiness design.

BunHonoPostgreSQLPrismaRedisBullMQNext.js 15
  • Structured four coordinated application surfaces: a Bun API and worker system, a Next.js web interface, a native iOS app, and a native Android app.
  • Implemented social photo workflows with uploads, privacy-aware location handling, AI image analysis, realtime feed updates, comments, reactions, bookmarks, private messaging, notifications, moderation, statistics, events, and account-data flows.

A multi-tenant SaaS case study for device-repair shops: a Rust and Axum API with schema-per-tenant PostgreSQL isolation, an Astro and Vue staff app, a QR and PIN customer status page with repair approvals, warranty certificates, realtime ticket updates, and a CI-gated Railway release pipeline.

RustAxumsqlxPostgreSQLRedisAstroVue 3
  • Built a multi-tenant repair-shop SaaS where every shop lives in its own PostgreSQL schema, with transaction-scoped tenant routing and an isolation test suite proving that data, sequences, and queries never cross between shops.
  • Implemented the repair workflow end to end: intake with printed QR slips, a staged repair pipeline, per-item customer approvals on a no-PII status page, handover rules, versioned warranty certificates, inventory, and owner reporting across 14 locales.

A Python and FastAPI matching service case study used by Fusio(n) for researcher-to-researcher matching, combining transformer embeddings, structured profile categories, bidirectional scoring, caching, and accelerated inference.

PythonFastAPIPyTorchTransformersSPECTER2all-MiniLM-L6-v2SciBERT
  • Implemented a companion ML matching service that Fusio(n) can call for collaborator ranking and workgroup formation instead of keeping all matching logic inside Laravel.
  • Implemented multiple matching paths: legacy SciBERT matching, SPECTER2 technical matching, fast structured category matching, and all-MiniLM-L6-v2 broad matching.

A Laravel and Inertia platform case study for research collaboration discovery, AI-assisted matching, team formation, real-time communication, and production-readiness design.

Laravel 12PHP 8.3React 19TypeScriptInertia.jsPostgreSQLpgvector
  • Mapped the technical architecture for researcher onboarding, profile-based discovery, semantic matching, workgroup creation, team workflows, and collaboration communication.
  • Implemented AI-backed matching patterns with provider abstraction, embeddings, pgvector similarity search, hybrid scoring, and reliability controls.

A research-analysis PWA case study that uses Rust compiled to WebAssembly for browser-side academic and policy data processing, with protected React workflows, offline persistence, and exports.

RustWebAssemblywasm-bindgenReact 19TypeScriptVitePWA
  • Built the full Litoscope Academic and Litoscope Policy workflows for processing exported data source files and policy Excel workbooks directly in the browser.
  • Implemented Rust/WASM data pipelines for parsing, citation shaping, policy-publication matching, usage metrics, date normalization, and generated intermediate research outputs.