A room from 1999. Files from today.
I build applied AI products at AUGE in Zürich.
Before that, I founded Pitcher, grew it to eight-figure ARR and sold it in 2022.
This is my room at the turn of the millennium. Open a document to see what I’m up to now.
Read my storySmall products, rooted in my own life.
On the workbench: Time Sloths and Boxing Stash.
Open the full portfolioMy father introduced me to computers with AutoCAD when I was seven. Both my parents are engineers.
Ankara, computer engineering at Istanbul Technical University, then Zürich. Computational neuroscience and neuromorphic engineering at UZH and ETH Zürich.
In 2007, I reverse-engineered the original iPhone and built Crossy, before the SDK or the App Store existed.
I founded Pitcher in 2010, sold it in 2022, and stayed until 2025. Now I’m building AUGE.
The Pitcher yearsFamily. Training since I was 16. Metal, and trying to understand the vocals. Questions that don’t have tidy answers.
Neuromancer. Snow Crash. Do Androids Dream of Electric Sheep?
eXistenZ, The Thirteenth Floor, and The Matrix. Three good reasons to question reality.
A Lamborghini Diablo and Michael Schumacher. Some things don’t need explaining.
A little more about meA hard problem. A small team. Something worth shipping.
Zürich, Switzerland
Mert Yentur / An ongoing exhibition
Computers. Brains. Companies.
Things I couldn’t leave as ideas.
Both my parents are engineers. My father introduced me to computers with AutoCAD when I was seven.
Apparently, crayons were too straightforward.
Neuromancer on the desk. A Lamborghini Diablo and Michael Schumacher on the wall. The Matrix, eXistenZ and The Thirteenth Floor waiting to be watched.
Brand-new books and VCDs. A Windows 98 computer. Plenty to be curious about.
I reverse-engineered the original iPhone and built Crossy: a snow globe driven by the accelerometer, one of the first native iPhone apps.
A tiny app. A big discovery: I could make something happen on a device that wasn’t supposed to be open yet.
Try shaking the exhibit.
After my computer engineering degree at Istanbul Technical University, I went to UZH and ETH Zürich to study computational neuroscience and neuromorphic engineering.
Silicon retina. Silicon cochlea. Experiments with a foveal vision sensor. I wanted to understand how brains work—and what we could make with that understanding.
I founded Pitcher and built its original prototype. A presentation tool grew into an enterprise platform with offline CRM and field workflows.
I was the founder, CEO, salesperson and an engineer on the product. The customer conversation and the code were parts of the same job.
The meeting could happen anywhere.
In-person meetings stopped. A field-sales platform needed another way to reach people.
I built the Node.js backend and integrations for remote detailing in approximately a month. Browser access and WhatsApp invitations took the experience beyond the sales representative’s iPad.
I grew Pitcher to eight-figure ARR and sold it in 2022.
I stayed on, first as CEO of a 150-person team, then as Chief Innovation Officer until 2025. As the team rebuilt the core platform, I returned to the prototypes that made the next version tangible.
2010–2025 / The Pitcher yearsLyrics I can see. Training data my own agent can use. A better video feed for my son. Private podcasts in the car. Applied AI, rooted in my own life.
I’m going deeper into Python, PyTorch and language models. Time Sloths and Boxing Stash are on the workbench.
The tools keep changing. The satisfaction of making something work doesn’t.
Let’s make somethingI’m Mert. I turn difficult ideas
into products people can use.
I turn open-ended problems into working products.
I work across problem definition, system architecture and rapid prototyping, then keep going through integration, debugging and launch. I’m especially useful when the direction isn’t obvious yet.
That’s the thread from reverse-engineering the iPhone in 2007, through Pitcher’s first iPad prototype, to the products I ship at AUGE. Sometimes I build directly; sometimes I direct AI agents or work with a production team. I own the architecture and the decisions, and I’m responsible for making the result useful.
Turn an interesting idea into a specific question. Decide what would make the answer useful.
Choose the architecture. Make the difficult part work in a prototype people can try.
Follow the failure through the system. Debug the assumptions as well as the implementation.
Get the product into people’s hands. Use what you learn to improve it and decide what comes next.
I’ve built this way at Pitcher and AUGE. Now I’m extending it into model work, going deeper into Python, PyTorch and LLM training.
Turn a question about data, architecture or a training objective into a small experiment. Connect the data pipeline to the training process, establish a baseline, and make failures understandable before a larger run.
Start with a behavior that needs to improve. Define examples and evaluation criteria, prototype a focused fine-tuning or feedback experiment, and compare it with the starting model on held-out cases.
Connect models, data, interfaces and real-world constraints into something people can use. Work through state, latency, context and failure modes until the product works reliably.
Most of these started as something
I wanted in my own life.
Currently at AUGE
I like metal. I like singing along. Understanding the vocals is a separate engineering problem. Beyond Chorus puts the lyrics where I can see them.
I’ve trained since I was 16. NAFA keeps logging and tracking fast, then gives your chosen agent a structured picture of your progress. Your agent handles the interpretation and planning.
Built for my son. Used by me too. Blocking fights the feed and loses, so Brain-feed curates instead: you pick the channels, and the feed shows nothing else.
Roadcast brings private subscription feeds to the car browser. Open it. Add a feed. Drive with your queue.
Explore Roadcast
The next things
I wanted to exist.
A family day planner in ten-minute blocks.
IN DEVELOPMENTA memory for where you put your stuff.
IN DEVELOPMENTI founded Pitcher in 2010 and built its first prototype. I was CEO for eleven years, its lead salesperson and an engineer on the product. I grew the company to eight-figure ARR and sold it in 2022. I stayed on as CEO of a 150-person team, then as Chief Innovation Officer until 2025.
I built and maintained the iOS app, the conversion engine and the analytics engine for most of that run. When the engineering team rebuilt the core platform for Pitcher NG, I returned to building the next version from the ground up.
Offline presentations. CRM. Remote detailing. Then a new platform. Fifteen years of turning the next difficult problem into something tangible.
Build the
first prototype.
Move the meeting
into the browser.
Sold at
eight-figure ARR.
Reimagine the
platform underneath.
I wanted to understand how brains work. And, somewhere along the way, what consciousness is. Apparently, building software wasn’t enough of a rabbit hole.
That took me to the Institute of Neuroinformatics in Zürich in 2008, to study computational neuroscience and neuromorphic engineering: electronic systems that borrow ideas from nervous systems, down to the way circuits sense and communicate.
I explored silicon retina and silicon cochlea systems. My project work included experiments with a foveal vision sensor, looking for practical ways to observe people and vehicles. The question was already: what can we make with this?
In an event-based retina such as the Dynamic Vision Sensor, individual pixels report local changes in brightness as they happen. Each event carries a location and a time. A moving edge becomes a stream of activity, without waiting for the next full camera frame.
A silicon cochlea takes inspiration from the inner ear, separating sound into frequency channels. Event-based versions encode activity as precisely timed spikes. Sight and sound become signals that other circuits can process as events arrive.
Back then, the practical applications were still years away. I completed the coursework, then chose building over finishing the degree: I wanted to make things people could use. Pitcher came next.
The fascination with perception stayed. AUGE means “eye” in German. Today, that curiosity shows up in a different form: lyrics in smart glasses, software that connects what you hear with what you see, information that belongs in the moment. Different technology. A familiar itch.
Back to what I’m building nowThat’s the number I want to optimize.
Both my parents are engineers. My father introduced me to computers with AutoCAD. I was seven. Apparently, crayons were too straightforward.
I drew talking-animal cartoons, made a magazine and handed it out at school with bubblegum attached. My father printed the copies at his office. The distribution strategy was edible.
Ankara, a computer engineering degree at Istanbul Technical University, then Zürich. WebObjects, an iPod simulator, then the iPhone: in November 2007, I reverse-engineered it before the SDK or the App Store existed and built Crossy, a snow globe app driven by the accelerometer, one of the first native iPhone apps. Silicon retina experiments, then Pitcher. The tools kept changing. The satisfaction of making something work didn’t.
After fifteen years at Pitcher, I’m building AUGE around problems in my own life. I choose the problem, decide how the product should behave, direct the work and keep pushing until it works that way.
I still like questions that don’t have tidy answers. I’m learning to build a small language model. Apparently, even my downtime needs an unresolved problem.
Making Engineering Ridiculously Tangible.
A hard problem. A small team. Something worth shipping.
Let’s build