Ship it and hope.
Merge the happy path, watch the store rating drift down, and file the crash reports under noise.
Senior Software Developer — React · React Native · .NET
I build web and mobile products end to end — the interface, the services behind it, and the performance work that decides whether either one survives contact with real users.
// bullet time
// incoming
Features are the easy part. What decides whether a product survives is how it behaves on a three-year-old phone, on a train, with 3% battery and a flaky connection. That is the work I actually do.
// resolved
Dodging these is the job. Not a side quest.
// the choice
Merge the happy path, watch the store rating drift down, and file the crash reports under noise.
Profile the render, read the native trace, reproduce it on the cheapest device you own, then fix the cause instead of the symptom.
// whoami

I'm a senior software developer in İstanbul. I work across React, React Native and Vue.js on the front, and C# / ASP.NET Core on the back — API design, business logic and the data layer included.
Most of what I've shipped lives where the hard parts are: realtime features, payment and subscription flows, map-based services, and the enterprise admin panels nobody demos but everybody depends on. For three years that meant health technology, building clinical products used across twelve countries.
I care about the unglamorous half — frame timing, cold start, memory, what happens when the network isn't there — because that is the half users actually feel. And I own the whole line: from the idea to the thing running in production.
// arsenal
// experience
08/2025 — present
08/2022 — 07/2025
09/2021 — 08/2022
12/2019 — 09/2021
07/2019 — 12/2019
// background
// case study
A social and commerce platform for car and motorcycle owners: a digital garage for your vehicles, a feed and forum for the community, maintenance and expense tracking, drive recording, realtime messaging, map-based service discovery and a marketplace for shops.
React Native · iOS & Android
392
commits
2
platforms shipped
60fps
scroll target held
62
E2E flows automated
Gated 230 console calls behind __DEV__, moved serialisation out of render bodies, and killed idle re-renders that were driving 60fps redraws on a static screen. Verified with gfxinfo frame capture on a real device, not a simulator.
Patched react-native-maps so images inside markers actually load, wrote Nitro module bridges, and fixed an Android TextInput state chain that leaked on every keystroke and overflowed the stack on teardown.
62 Maestro end-to-end flows covering auth, garage, feed, social, shop mode and account. A single QA pass across them surfaced 178 fixes.
Redux with Saga and persisted state, MMKV for hot storage, SignalR for realtime, Firebase for auth and messaging, In-App Purchase for subscriptions, Sentry for whatever still gets through.
// stack
// selected work
Health platform running in twelve countries: patient-to-clinic matching, automated report flows and advanced filtering over clinical data.
AI-assisted clinical trial discovery and oncology data analysis for health professionals. Presented at ASCO 2024 and 2025.
Corporate web platform. React front end, .NET backend services and the REST layer between them; responsive structure and data flows handled end to end.
Loyalty and reward management platform: points, balances, campaigns, transaction flows, validation and the admin panel that operates all of it.
Modular platform joining social features to a reward system: point transfer, shipment tracking and payment integrations.
Industrial tracking for energy and natural gas facilities. Expo-based field application with realtime sync between field teams and backend.
GIS-integrated real estate information system: map-based, realtime property and asset tracking.
Location-based social network built on geo-fencing and map services, with clubs and a live feed.
Education ecosystem covering school administration, content sharing and teacher-to-parent communication.
// signal
For three years my work sat next to applied AI in oncology — products that read clinical data and put an answer in front of a doctor. That is a domain where a plausible-looking wrong answer is not a bug report, it is a patient.
Dr-Arturo AI and Ask Fiona AI help health professionals find clinical trials and make sense of oncology data. My side was the product layer: turning model output into something a clinician can act on, and making the uncertainty legible instead of hidden.
SYNERGY-AI matches patients to trials across twelve countries. The hard part was never the matching — it was the filtering, the reporting and the data hygiene that make a match trustworthy.
I run large refactors, QA passes and migrations through coding agents, with tight and verifiable success criteria. The skill is context design and knowing exactly what to check. The bottleneck has moved from typing to judgement — telling a correct system from a plausible one is becoming the whole job.
// end of line
But there is an inbox. If you're building something that has to work on real devices, for real users, say hello.