Emir Can Manici

Full-StackSoftwareEngineer

Backend-focused engineer shipping production Python/FastAPI systems — real-time infrastructure, multi-provider AI routing, three products live in production.

3

Products live in production

131

FastAPI endpoints shipped

500ms

First AI token via SSE

18 mo

In production engineering

Open to Relocation & Visa Sponsorship
0
Products live in production
0
FastAPI endpoints shipped
0ms
First AI token via SSE
0 mo
In production engineering
Emir Can Manici

Emir Can Manici

Full-Stack Software Engineer

Çerkezköy, Tekirdağ, Turkey

About

I'm a backend-focused full-stack engineer with 1.5 years shipping production Python/FastAPI systems for US and Turkish clients, while finishing a B.Sc. in Information Technology. I own products end-to-end — schema design, API, deployment — not just the feature I was assigned.

At BLGN Studios I build backend features across a 131-endpoint FastAPI service powering Mira AI: chat, an in-browser IDE, image generation, a social feed. I built the realtime voice gateway (a WebSocket proxy fronting OpenAI Realtime and Google Gemini Live) and SSE token streaming that cut perceived response time 50–90%.

Before that, at Emrisoft (US, remote) I designed the FastAPI backend for a legal-document AI platform parsing PDF, Excel, Word, and Turkish government UDF files, with automatic failover across OpenRouter, OpenAI, and Gemini to survive provider outages.

Outside contract work, I solo-build and run two live products — MeetBook, a privacy-first location-aware book exchange with WebRTC voice/video, and Kuafor, a salon booking and management platform running for a real local business. Open to relocation, visa sponsorship, or full-remote.

Selected Work

RealProducts,ActuallyLive

Three systems I own end-to-end — from schema to deployment. Not tutorials, not clones. Click any screen to see it full-size.

Solo-built2025

MeetBook

Location-Aware Book Exchange

A privacy-first, location-aware platform for discovering and trading books nearby. Modular monolith: FastAPI, PostgreSQL + PostGIS, Redis, React Native/Expo, WebRTC.

FastAPI + PostgreSQL/PostGIS + Redis backend · React Native/Expo client · WebSocket chat + WebRTC voice/video · Argon2id auth with rotating refresh tokens

~82%

Calls connect P2P via WebRTC

30 days

Auto-deleted location history

1 hr TTL

Ephemeral TURN credentials

FastAPIPostgreSQLPostGISRedisReact NativeExpoWebRTCArgon2id
Solo-built2025

Kuafor

Salon Booking & Management Platform

A full booking-to-management platform for a real hair salon — public booking site plus an internal admin dashboard for staff scheduling, appointments, and WhatsApp notifications.

Next.js frontend · Express (Bun) backend · PostgreSQL + Redis · Self-hosted WhatsApp notifications · Nginx + Fail2Ban · Docker Compose

0

Server renders for cached public pages

5-step

Booking flow, zero page reloads

Isolated

WhatsApp notifications

Next.jsExpressBunPostgreSQLRedisNginxDocker
Contract work2025

Avukado'ya Sor — Legal AI Platform

AI Document Analysis

A legal-document AI platform for Turkish legal professionals — parses PDF, Excel, Word, and Turkish government UDF files, with a built-in document editor and multi-provider AI failover.

FastAPI backend · PostgreSQL + Redis · Automatic failover across OpenRouter, OpenAI, and Gemini · MCP context-orchestration layer for unstructured legal text

3

AI providers with automatic failover

4

Document formats parsed (PDF/XLS/DOC/UDF)

FastAPIPostgreSQLRedisOpenRouterMCPDocker

Experience

WhereI'veBuilt

Full-Stack Developer (Contract)

BLGN Studios

Dec 2025 – Present
  • Shipped backend features across a 131-endpoint FastAPI service (~6,500 LOC) powering chat, an in-browser IDE, image generation, and a social feed for Mira AI.
  • Built the realtime voice gateway — a WebSocket proxy fronting OpenAI Realtime and Google Gemini Live; deployed with Dokploy and Docker Compose.
  • Delivered SSE token streaming: first token in ~500ms, cutting perceived response time 50–90%.
  • Designed intent-based model routing with a sub-1ms regex fast path handling 5–15% of traffic at zero LLM cost, holding routing spend to ~$1–5/day at 10k requests.
  • Shipped iOS/Android features from a shared React Native/Expo codebase; added GitHub Actions CI (ruff, pyright, pytest + testcontainers on MariaDB) gating every PR.

AI Systems Engineer (Contract)

Emrisoft (United States)

Jun 2025 – Dec 2025
  • Designed the FastAPI backend for a legal-document AI platform parsing PDF, Excel, Word, and Turkish government UDF, with PostgreSQL and Redis on one cost-optimized Docker host.
  • Built automatic failover across OpenRouter, OpenAI, and Gemini to survive provider outages.
  • Built a Model Context Protocol (MCP) context-orchestration layer to improve retrieval on unstructured legal text; added a router agent and prompt-guard validation layer.

Web Developer (Contract)

İşkur — Turkish Employment Agency

Mar 2025 – Jun 2025
  • Built and shipped data-driven web apps (FastAPI, Next.js, Tailwind) for Türkiye's national employment agency.
  • Dockerized and deployed to AWS/GCP infrastructure.

Skills

TechnicalArsenal

Backend

FastAPINode.js / ExpressRESTWebSocketsWebRTCSSEJWT / OAuthSQLAlchemy / Alembic

AI & LLM Systems

OpenAIGeminiOpenRouter multi-provider routingAutomated failoverRAGModel Context Protocol (MCP)

Frontend & Mobile

ReactNext.jsReact NativeExpoTailwind CSSZustandTanStack QueryVite

Data & Infrastructure

PostgreSQL + PostGISRedisMariaDBDockerNginxGitHub ActionsAWS (EC2, S3)GCP

Languages & Testing

PythonTypeScriptJavaScriptSQLpytesttestcontainerspyrightruff

Credentials

Certifications&Training

AWS Certified Cloud PractitionerIn Progress
Docker and Kubernetes: The Complete Guide
Advanced React Patterns and Performance
Recruiter & CTO FAQ

Every Question
Answered Honestly

38 questions across 6 categories — built for HR recruiters, CTOs, and hiring managers. Every answer is specific, honest, and backed by shipped production systems.

0
Total Questions
0
For HR / Recruiters
0
For CTOs / Tech Leads
Browse by Topic
Top Questions — Start Here
HR & Hiring · 11 questions

Still have questions?

I reply to every email personally — usually within a few hours, always within 24.

LLM-ready:llm.txtllms.txt

38 questions · 6 categories · Designed for HR & technical hiring

Let's Build
Something Real

Open to remote roles, on-site relocation, and visa sponsorship worldwide.