Simon Doba
AI and DevOps Engineer
I build AI systems, cloud infrastructure, and web applications. Based in Berlin, I work at the intersection of engineering and design - from architecture to interface.
I started building software in 2017 at a small startup in Stuttgart - Android apps, route optimization, UI design. That's where I discovered I care equally about how something works and how it feels.
From there I moved to Berlin, studied Computer Science at TU Berlin, and worked as a working student at enterprise companies: MHP (Porsche subsidiary), T-Systems, and now Orbit. In between there were seven months as CTO of an early-stage venture, which is where I learned what a technical decision costs when nobody else is going to make it. Each role pulled me deeper into infrastructure, DevOps, and eventually AI.
At MHP, I ran a GenAI feasibility study for a DAX-10 corporation - LLM-powered documentation automation that hit 80% quality against a 50% gate. At T-Systems, I helped build the Digital Ventures department and shipped satellite teletext extraction systems. At Orbit I'm building a multi-LLM chatbot platform with RAG pipelines for production environments - and writing full-stack again, which is where I started: an OpenWebUI fork and internal AI-enabled tools in Next.js for a DAX-40 corporation.
On the side, I authored a research study analyzing 24,560 pull requests to understand how AI coding agents impact CI/CD pipelines. That research drives the consulting I do for teams adopting AI tools.
AI Consultant
Orbit GmbHBuilding a multi-LLM chatbot platform with RAG pipelines, vector databases, and modular middleware for production customer environments. Full-stack alongside it: an OpenWebUI fork and internal AI-enabled tools in Next.js for a DAX-40 corporation.
Software Engineer
T-Systems InternationalSatellite teletext extraction via UDP multicast streams. Co-built the Digital Ventures department. DevOps infrastructure with Terraform on OTC and AWS.
Backend Developer / DevOps
MHP (Porsche Subsidiary)GenAI feasibility study for a DAX-10 corporation: LLM-powered documentation automation. Azure cloud infrastructure with Terraform for DAX-40 companies.
CTO
Stealth Startup, BerlinEarly-stage venture building an AI-powered affiliate and gift discovery platform, monetised through affiliate links. Set the technical architecture and the product roadmap, and built the platform from nothing to a working early prototype. Designed the recommendation approach around a hard constraint: the recipient is never a user, so there is no purchase history to filter on. Recommendations came from the occasion, the relationship and what the giver knew about the person. Shaped product strategy, business model and go-to-market with the founding team.
Junior Fullstack Developer
Ascaion GmbHDynamic grant application portal for cultural organizations. Automated testing and multi-environment deployment pipelines.
Junior Software Engineer
aceArt GmbHAndroid app for route navigation with 100+ waypoints including route optimization. UI/UX design and dashboard development.
B.Sc. Computer Science
Technische Universität Berlin
2018 – 2026
Thesis: Design and prototypical development of an interdisciplinary data platform for harmonizing heterogeneous daylight data.
Certifications
- AWS Cloud Practitioner
- AI Certificate – Charité Berlin
Do AI Coding Agents Break Your CI/CD Pipeline?
An empirical study analyzing 24,560 pull requests across 447 GitHub repositories. Found that AI-authored PRs have 19.4% higher odds of CI failure, with small PRs failing at 2x the rate of human changes.
Read the full study