AI · Automation · Video
I build automations, AI video pipelines and live systems. I built all of it on my own work first: FeastFlows has real customers, and Resources has the step-by-step version of each piece. The whole road from bicycle courier to here is below too.
01How I built myself
I started out delivering on a bicycle. Today, through FeastFlows, I support restaurants in their digital transformation. Everything in between I taught myself: sports and sales, then e-commerce, social media, automation, and eventually writing code with AI and putting live systems into production.
I don't explain anything I haven't run myself. FeastFlows has real restaurants and real orders behind it, and the ordering system I built for Asil Su is still running. Every guide in Resources was installed on my own machine first, which is exactly where the lists of traps come from.
Who I am and what I built: the long version →02What I build for you
Two kinds of work come to me. The ones that repeat every week and wear you down, and the ideas that stall at “how do I even build this”. I build both, and I built all of it on my own work first.
Run a restaurant? FeastFlows is the right place.
If you do the same thing by hand every week, a system is missing there. From taking orders to following up with customers, from reporting to reminders, I write the code that turns repeating work into something that runs in the background. You see the result; the work runs itself.
A shoot means a crew, a location, equipment and a whole day. I build a line that produces the same ad from a short brief: script, scene and video out of one system. A shoot that costs thousands becomes a production that costs tens. The real difference is elsewhere though: you cannot run a shoot twice, but you can run a system again tomorrow with ten variations.
Follower count is not a business. What matters is how a follower becomes a customer. We take on the work you already do on social and wire it into one line: where the content leads, how an incoming message gets answered, and who ends up ordering. We turn attention that currently leaks away into revenue.
I build the system and hand it over working. You see the result.
I build it in your own accounts, in sync with you. You learn not just the system but how it was built: which piece connects to what, and where to look when something goes wrong. Next time a small error shows up, you can fix it yourself instead of waiting for me. What I write about in Resources, done on your work this time.
The second one runs in sync with you, so it is priced differently.
I listen to what you do and what you want off your plate. Together we pin down which job moves into the system.
Handed over, or built together with you. Your call.
The system goes live in your hands. We walk through it together once, end to end.
It stays running and I stay behind it. When something breaks or your work changes, you call me.
The calendar opens, you pick an open slot, and the Google Meet link arrives automatically.
The first call is free. We talk about what can be built and what it would cost.
03What I have built
Step by step from 2022 to today: everything I tried, learned, and built. The whole journey that led to my live projects.
Highlights
Full journey
I'm a graduate of Dokuz Eylül's Electronic Communication Technologies program; I had a technical grounding, but the electronics side wasn't for me. I'd chosen the sports sector and was a stranger to startup and project-development culture. A close friend from the sports scene brought the Atlet.IA idea, and that was my first step into this world: I contributed my sales, marketing, and sports-industry experience, and we entered the TÜBİTAK 1512 (BiGG) process together. My first time working with an engineering team and seeing a real software project from the inside.
I wanted to build a digital business and a location-independent, online life. I set out with dropshipping training from Gürkan Zone: this is where I learned the fundamentals of building a business, from Shopify sites and Meta ads to competitor analysis, budgeting, and copywriting. To cover the training, my UK company, and my ad budget, I took on extra work including courier jobs. Dropshipping wasn't for me, but the copywriting and social-media fundamentals I gained carried me into social media a few months later.
I changed course and started producing content on social media and building my own business in sports. What I carried over from e-commerce (building websites, competitor and ad analysis, shooting and editing video) paid off here; I began producing sports content in English. To cut costs I learned to use Hostinger instead of Shopify at a tenth of the price, and set up a website where I could prepare and deliver PT (personal training) programs to my first paying clients.
Looking for a way to cut costs even further, my developer friends pointed me to HTML, CSS, and JavaScript. So I coded my first own website (HealthyLiving) with the help of Copilot and shipped it with Firebase: a humble first project, fully made with my own hands, where I offered my PT and nutrition programs. This is exactly where my journey of building with AI began.
When my first social-media attempts didn't get the results I wanted, I earned Meta's six-course professional Social Media Marketing certificate (Coursera) to learn the craft properly. In the same period, I was working as a Yemeksepeti courier to raise the capital I needed to offer services in sports. On one delivery, a small coincidence at a door led me to my first social-media management job, and then to the long journey with Gözde Biber that I'm still on today.
My first taste of automation was ManyChat; I built DM flows to grow influencer accounts and turn followers into customers. In the same period I took over a simple QR membership-lookup app a developer friend at the gym had built with Google Apps Script and turned it into a real system: a trainer enters a member's photos and measurements into a Google Form, the data lands in a table, and the member scans the gym's QR and enters their phone number on the page to compare their own history (e.g. September vs October). This was my first real contact with backend, databases, and system design.
Discovering that backend existed sparked my curiosity; researching on YouTube, Burhan Kocabıyık's videos introduced me to an approach: instead of learning Python or JavaScript, learn n8n and build the automations you need far faster. 'Creating something from nothing' pulled me in, and I started learning n8n. My first goal was to solve my own work: producing content for the gym. I'd give the automation a topic; using Google Gemini it searched PubMed for scientific papers, picked the best ones, and generated Instagram carousel (slide) content from them. This was my first no-code n8n project.
For my father's water distributorship (Asil Su) I built a real, backend-driven web app with React and Firebase: a customer storefront, an admin panel, and a field operator panel in one codebase. My first end-to-end full-stack project, still live. The Agora factory invited me to discuss rolling it out to all their distributors.
I'd learned n8n from his videos back in 2024, then filled out a form to join his automation team. Out of many applicants, in October 2025 Burhan Kocabıyık, one of Türkiye's leading AI educators, personally reached out; we met, and I started working on his team.
I got into AI video by producing UGC-style ad videos for apps like Prototipal and Mixy: short promos that introduce the app, with no expensive shoot. Along the way I learned tools like fal.ai, produced the same work in different languages, and it ran in many countries including Spain, the US, and Turkey.
I turned my ad-video work into a pipeline: a system that takes a short brief and, with character and rules from a NotebookLM vault, generates Kling 3.0-ready 8-scene UGC storyboards. I used it across sectors, from dental clinics to restaurants.
Real vs AI: the same ad
A series of experiments with pipelines that generate video from code, from auto-producing script, visuals, and voiceover from a single input to motion graphics. The goal was scalable, repeatable video systems for YouTube.
At GoatStarter, a selective accelerator where founders build AI-first businesses from scratch, I worked on the screening side: running intro calls with candidate founders, assessing their projects and readiness, and onboarding the right ones into the program.
The agency I founded to support restaurants in their digital transformation. The first problem it solves: bringing back a customer who came once. With a WhatsApp-based loyalty system, automated win-back campaigns, a QR menu, and a website, I give restaurants not just a digital storefront but returning customers every month. Live now, and I keep growing it.
04So you can build it too
Every tool, repo, and guide behind my posts, with installation steps included. There's no room to explain it on Instagram; there is here. Type the code from the video, or browse by topic.
The daily posts go on Instagram; the written version, with setup steps and the traps, lives here.
Instant access by code · no signup
OmniRoute hands your work to another model when you hit a limit. The two free providers its README suggests are flagged "avoid" in its own table
Latest →
05After the setup
The guides get a tool installed. The hard part starts after that: which part of your work to automate, where to begin, and who to ask when something breaks. I put together a page about the community I am part of.
Inside there is a short introduction video and two different ways to join.
06Let's build yours
The fastest route is picking a slot on the calendar. If you would rather write, email works too.
If you write, I usually reply within 1-2 business days.