Solo engineer with four apps live on the App Store this year. Production-grade iOS and AI integration, end to end — design, build, ship, recover from Apple rejections, repeat.
Real apps. Designed, built, deployed, and recovered from Apple rejections — solo, end to end. The four on the App Store are the proof.
An English-learning app that translates in five distinct voices using Claude Haiku — formal, casual, poetic, direct, conversational. Offline WordNet + cmudict for vocabulary. Always-on Fly.io backend. Recovered from Apple Guideline 4.2 rejection on the first resubmit.
View on App StorePrivacy-first social network with Dunbar-7 close-circle model. Zero trackers, zero ads. KVKK + GDPR compliant.
View on App StoreClinical decision support for hair restoration + 5 medical branches. Deterministic mathematics — no LLM in the critical path. KVKK + GDPR ready.
Visit siteiOS native OBD-II diagnostics with a digital intervention layer. Mode 04/01/08, CoreBluetooth ELM327, five-hypothesis log-odds engine with anti-bypass invariant — diagnostic results are timestamp-anchored.
TestFlight on requestMeditative wallpaper generator with native iOS rendering.
App StoreAI Chat & Flirt — Flokrater protocol with a custom signal layer.
App StoreEach one was proven on my own production apps before it became a service. No "I read a tutorial" — these are battle-tested patterns.
Guideline 4.2, 2.1.0, 1.2.0, ITMS errors. Resolution Center replies that read like a reviewer wrote them. 24–72h turnaround.
Solved on Lingua Nokta (4.2 iPad layout) and Aura Social (1.2 UGC moderation) — both live now.
Anthropic Haiku, Sonnet, Opus. Production pipelines with prompt caching, RAG, memory, cost optimization. iOS and cloud both sides.
Lingua Nokta runs Claude Haiku live, five-tone translation at ~$3/mo operational cost.
Swift + SwiftUI from spec to TestFlight in one week. Apple Sign-In, Supabase backend, push, iPad universal, App Store ready.
Two apps shipped this way in May 2026 — same pipeline available now.
Python + FastAPI + Supabase + Fly.io. Always-on services, prompt caching, RLS-secured databases, Realtime sockets, Cloudflare Pages distribution.
Lingua API runs on Fly.io always-on, Aura Social on Supabase with full RLS — both production-grade.
Deterministic computing cores, mathematical engines, binary protocols. OBD-II ELM327 stacks, CoreBluetooth low-level, custom runtime substrates for select work.
Motor372 runs full OBD-II Mode 04/01/08 with log-odds + anti-bypass invariant — production-tested.
Three.js + custom GLSL shaders + post-processing. Particle systems, real-time rendering, mobile-first responsive. From landing pages to fully interactive 3D experiences.
Built this site, plus interactive WebGL games with bloom post-processing and custom shader pipelines.
Not a list of things I've heard of. Production-tested, currently running.
A production runtime I've been building since 2024. Names are public — internals (architecture, formulas, parameters) stay under NDA. Licensing available for select partners.
The core production engine. An intent gate, a multi-tone language engine, a signal layer and a decision pipeline — all running over a deterministic mathematical substrate. Lingua's translation, Aura's matching and Akdağ's clinical scoring all pass through here.
Medical decision support layered on Yıldız. Healing-window dynamics, dose-response models, Bayesian update across branch atlases. The clinical core inside Akdağ Medikal — deterministic, no LLM in the decision path.
The command center. Multi-agent coordination, parallel production pipelines, autonomous output gating. Holds Yıldız + Asa and their siblings under one address and one process tree.
Plugin framework. Domain modules — medical, automotive, language, photo, AR — plug into Yıldız through a uniform call surface. Every new capability is a new Cip plug, not a fork of the core.
Names public, internals private. For licensing, technical due diligence or partnership conversations — NDA first, technical details after.
I'm Muhammet Barış Akdağ, an indie iOS and AI engineer based in Istanbul. I write Swift and Python, design in Figma, deploy to Fly.io and Supabase, and ship to the App Store every few weeks.
No traditional CS degree — I'm finishing a Bachelor's in Psychology. I learned engineering by shipping things: four App Store live, two more in production, infrastructure built solo from runtime to client.
Email is fastest. I reply within 4 hours during GMT+3 business hours, 12 hours otherwise. Happy to do a 15-minute call before any commitment.