Tuğra Gülsoy
Co-Founder
International Mathematics Olympiad competitor and CS background. Studying at TU Munich.
About CellWise
We make other people's files fit your format.
CellWise takes the data you receive — in whatever shape it arrives — and turns it into one consistent, validated format you can actually work with.
Mission
Stop reformatting spreadsheets by hand.
CellWise turns a recurring manual chore into a workflow you set up once: AI-assisted reading and column matching, a review step you control, validation against your own rules, and a deterministic transform that produces the same output every time.
Core insight
The data is usually fine. The formats are not.
The numbers a client sends are rarely wrong. They are just laid out differently, named differently, and formatted differently from the last file — and from what your system expects. That gap is where the hours go.
Market context
Built for anyone who receives data they did not format.
Every organization that receives data from outside — clients, suppliers, acquired companies, systems nobody has replaced yet — receives it in a format somebody else chose.
Making those formats agree is manual, repetitive, and almost never written down. When the person who knew the rules leaves, the rules leave with them.
This is not a small-company problem or a large-company problem. It scales with the number of sources you deal with, and it never gets smaller on its own.
Team
Four co-founders, one mission.
Emre Güner
Co-Founder & CBO
Workup and YGA alumnus, bridging customer needs and business strategy. Studying at ETH Zurich.
Arda Gürses
Co-Founder
TÜBİTAK winner, focused on engineering and practical AI for the teams that work with the data. Studying at TU Munich.
Emre Bora Kaynar
Co-Founder
TÜBİTAK winner and two-time robotics world champion. Product architecture and AI engineering. Studying at TU Munich.
Values
The principles we build against.
One job, done properly
We would rather be the best tool for cleaning and standardizing files than an adequate tool for everything. The roadmap is narrow on purpose.
Reproducible over impressive
Same input, same rules, same output — today and next quarter. A result you cannot reproduce is not a result you can defend to a client.
You can see the work
Every mapping, correction, and rejected row is visible and exportable. Nothing happens in a step you are not allowed to look inside.
AI where it helps, rules where it matters
Models are good at reading ambiguity. They are the wrong tool for producing output that somebody has to sign their name under.
Your data stays yours
Customer files are not used to train models. Data isolation and ownership are product principles, not settings.