10 teams, 60+ attendees, 24 hours, 10 amazing projects, 4 tech jurors from eSolutions & Tchibo, and loads of brainstorming - the 10th edition of the eSol x Tchibo Hackathon in a nutshell.
Another year, another milestone! This year’s Hackathon marked a full decade of eSolutions hackathons and the second edition together with our friends at Tchibo. Under the theme Glitch the System, colleagues from Bucharest and Hamburg teamed up across eSolutions, Tchibo, and mixed squads to turn bold ideas into working prototypes. We were very excited to see what everyone came up with, and we’re sure you are too, so let’s take a look at how it all went on 18–19 September!
Overview
The eSolutions x Tchibo Hackathon 2026 was all about pushing boundaries, collaborating across organisations, and unleashing creativity. With the anniversary theme Glitch the System, participants were encouraged to rethink everyday workflows, retail experiences, accessibility, and internal tools, as long as their ideas were innovative and techy. Over an intense 24-hour period, teams worked tirelessly to bring their visions to life, culminating in dynamic presentations, live demos, and insightful Q&A sessions.
The Teams & Their Projects
With the clock ticking and energy running high, ten talented teams dove headfirst into the challenge, each bringing a unique perspective and skill set to the table. Let's meet them!
Team THE DINOSAURS
Project Description: Berkley is an AI loyalty assistant designed for both shoppers and loyalty managers. On the customer side, it helps maximise loyalty benefits by suggesting basket tweaks, promotions, and category combinations that unlock more points. On the company side, it stress-tests promotions against large volumes of real baskets to surface edge cases and potential exploits, then recommends safer rule adjustments. The solution is built to integrate with any loyalty platform, not just a single retailer.
Team Members: Cezar Alin Ene, Gabriel Petre, Andrei Lipan
Team BREWCHAREST 2.0
Project Description: A Product Studio suite, including an iOS app, that takes a retailer from holding a physical product to a complete product detail page in about a minute. Users capture three photos plus a LiDAR / 3D scan; the platform then generates measurements, product copy, enriched media, and shop-ready structured content. Extra supplier files (Excel, PDF, invoices) can be uploaded to refine results, making it useful for retail backends, marketplaces, and second-hand flows alike.
Team Members: Joris Raphael Uecker, Philipp Schumacher, Fjodor Scharmacher, Mark Chiricenco
Team RUNTIME TERRORS
Project Description: GitWrapped turns 12 months of Git activity into a clear yearly review - stories and numbers, without scores or ranking. Developers connect Git sources with read-only access; AI aggregates commits, line changes, merge requests, and project signals into a shared report that both developers and HR/managers can see. The goal: replace incomplete tickets and “last month bias” with a full-year picture across repositories.
Team Members: Eva Simpetru, Ionela Baltaretu, Robert Tatuc, Larisa Biriboaca
Team TYPESHYT
Project Description: An AI-powered hackathon planning CMS that turns one prompt (plus optional guideline files) into a full event setup: title, description, schedule, guidelines, and participant-facing pages. Participants can create or join password-protected teams, while organisers keep a single source of truth for logistics. The team also added screen-recording demos so finished projects are easier to share after the event.
Team Members: Daniel Reimus, Maximilian Kanurski, Liam Eim, Alexander Khuu
Team THE CONTAINERIZED RACOONS
Project Description: An agentic onboarding platform that turns wiki-style onboarding pages into role-based checklists for new employees or clients. Deterministic steps handle the predictable work; an AI agent interprets responses, follows up on pending requests, makes access decisions (e.g. junior vs. senior), and adapts the plan when humans give new instructions. A buddy chatbot lets newcomers track progress and ask questions via QR / URL.
Team Members: Ionut Popa, Mihai Oprescu, Traian Rusu
Team POSSIBLE GLITCH
Project Description: Business Demand Helper is a clickable wizard that helps non-technical store and business users describe problems in natural language and leave with a complete, backlog-ready ticket. The tool pre-fills Jira-like fields from context, asks only the missing questions in plain language, and aims to cut clarification loops on a very high ticket volume — with a clear path toward future guided resolution.
Team Members: Timo Boos, Alex Constantin
Team SPRAY N' PRAY
Project Description: Can I Buy This? is an AI shopper that tests real e-commerce accessibility by simulating disabled customers (keyboard-only, mouse-only, screen-reader / blind scenarios) trying to complete a purchase - without being told how the shop works. It navigates via screenshots rather than DOM cheating, then reports keystrokes, decisions, timing, and whether the mission succeeded, so teams can answer a practical question: can a disabled customer actually buy from the web shop?
Team Members: Robert Milcu, Ngoc Hien Le
Team BEAN THERE, CODED THAT
Project Description: Meet Owie is an embeddable in-app support assistant. When something breaks in an internal tool, users chat about the issue; the assistant can capture screenshots, console logs, and network diagnostics, then draft a developer-ready Jira ticket with evidence attached. Packaged as an npm module with setup docs for AI agents, it can be dropped into Tchibo or eSolutions apps with minimal effort.
Team Members: Jona Bihl, Alexandra Sebold, Samuel Riascos, Victor Kuzu
Team ONE DIRECTION
Project Description: Fridge Police tackles the classic office fridge problem without creepy always-on surveillance. Door sensors trigger recording; cameras and AI detect who opened the fridge and what went in or out. “Thieves” land on a public wall of shame (office TV), stock per person is tracked, open-door alerts help prevent spoilage, and unknown faces can be labelled in an admin panel.
Team Members: Cristina Andries, Florentina Bobocescu, Alin Jderu, Marius Craus
And last, but not least...
Team CATPUCCINO
Project Description: Catpuccino OS is an agentic operating system concept for businesses - a flexible, white-label foundation with a polished UX shell (window management, themes, spotlight-style search) plus practical modules. Highlights include a self-improving knowledge base companion (the team’s 3D cat mascot), on-call / incident assistance with suggested fixes, and an experience designed to keep employees engaged while tackling real operational problems.
Team Members: Florentina Bibiloiu, Alexandru Opran, Maximilian Müller
And the Winners Are…
? First place: Brewcharest 2.0
? Second place: THE DINOSAURS
? Third place: POSSIBLE GLITCH
Thank you all for joining, to our partners at Tchibo Romania, and to everyone who showed up to support the teams. Here’s to a decade of experimentation and to whatever we build next!
Check out our cool video!
RO
EN
