English · Deutsch
Full-Stack Developer · AI-Assisted Product Builder
My background is in robotics and camera-based quality inspection. At ZF, I built lab prototypes that combined robots, cameras and inspection software. Today I work on web applications and automation, using AI tools throughout development.
These eleven applications run locally. Each project page includes screenshots, the reasoning behind the implementation and test results.
| Project | What it does | More detail |
|---|---|---|
| Scope Quant | Replays trading simulations with visible order decisions, shared risk controls and chronological strategy comparisons. | Accounting, timing and research checks |
| Scope Pulse | Imports social media metrics, compares posts and prepares reports with reviewed suggestions from a local model. | Calculations, imports and model checks |
| Scope Clips | Turns spoken videos into short clips with local transcription, editable captions and MP4 exports. | Media, model and browser checks |
| Scope Trace | Follows material lots through parts, assemblies and shipments. Quality cases retain holds, rechecks and release decisions. | Genealogy and workflow checks |
| Scope Dispatch | Plans technical service jobs around skills, arrival windows and travel times. A local model prepares requests for review. | Scheduling comparisons and checks |
| Scope Cell | Simulates a UR5e inspection cell with visible movement commands, part tracking and recovery from faults. | Motion, geometry and process checks |
| Scope Assist | Answers questions about selected technical documents with a local language model and opens the quoted passage. | Model evaluation and limits |
| Scope Vision | Finds scratches and missing holes in generated images of metal plates, then records a person's review. | Model evaluation and tests |
| Scope CRM | Keeps customer assignments, appointments and notes together. CSV imports are checked before they are saved. | Roles, import and browser tests |
| Scope Audit | Checks a page's HTML and internal links, explains the findings and saves the report. | Checks and limitations |
| Scope Monitor | Checks HTTP services at regular intervals and keeps a record of failures and recovery. | Monitoring rules and tests |
I developed these projects with Codex; the case studies explain my role and how AI was used. Only the presentations are public. Application source, runtime files, databases, models, training and evaluation data, original documents and recordings remain private. Private schedules, search indexes and exported dossiers are not included. No open-source license is granted for the original portfolio material. The examples use synthetic demonstration material; the results come from local tests.
I also developed an application with AI assistance to prepare media and captions, schedule posts for two X accounts and keep track of publication results. The case study explains how it handles uncertain API responses, repeated preparation failures and the separation of test posts from business analytics.
As of 19 October 2026, image publishing is active. I am still completing the video extension. Source code, account details and operating data remain private. Project status and reported tests.
Data Integrity Toolkit checks whether research exports are complete and internally consistent. It flags missing or contradictory information instead of filling in the gaps.
This table describes its behavior; it is not a screenshot of the application or a sample of private research data.
| Input | Result |
|---|---|
| Complete and consistent | Verification result |
| Changed or contradictory | Rejection or a description of the inconsistency |
| Incomplete | Missing information is identified; no replacement values are generated |
The tool was developed with AI assistance. Further integration tests are still in preparation. Its source, datasets, domain rules, internal test material and implementation details remain private.
Software Developer · KTC Saarbrücken · Research & Development · 2021–2024
I worked on manufacturing and assembly processes, built lab prototypes, and configured robots, vision software and operator interfaces.
One project was an inspection prototype that I built and configured using a Universal Robots robot and a SensoPart VISOR camera. The robot took machined mechatronic plates from the production line, presented them to the camera at defined angles and sorted them by inspection result. I also prepared technical drawings for parts made by the in-house toolmaking team.
The work included training image-based inspection models, adjusting grayscale tolerances and compensating for lighting changes across three shifts. After presentation and refinement, our department gradually introduced the solution across 12 production lines. This helped standardize quality assessment and reduce labor and scrap costs.
I designed, developed and deployed a web-based CRM with customer management, appointments, user roles and data import, and integrated it into the client's operations.
The project is complete. The client's code and business data are confidential. This portfolio describes the work I delivered; the system's current production status has not been checked.
| Area | Experience |
|---|---|
| Robotics & machine vision | Universal Robots · SensoPart VISOR · Camera-based inspection · Image-based model training |
| Web development | PHP · Laravel · JavaScript · TypeScript · React · Node.js · SQL · Git |
| Computer vision portfolio | Python · PyTorch · Pretrained image features · Anomaly detection · Held-out model evaluation |
| Local document AI | Document retrieval · Local language models · Source references · Regression evaluation |
| Service planning portfolio | Constraint-based scheduling · Route planning · Local AI intake · Transactional plan updates |
| Manufacturing traceability portfolio | Material relationships · Quality cases · Event history · Checked imports |
| Media processing portfolio | Local speech recognition · Transcript navigation · FFmpeg · Caption rendering · Versioned exports |
| Social media analysis portfolio | Checked CSV imports · Cumulative measurements · Weighted metrics · Local model drafts · PDF reporting |
| Trading research portfolio | Event-based simulation · Shared cash ledger · Market replay · Chronological validation · Explicit execution assumptions |
| Engineering | Process analysis · Prototyping · Workflow automation · APIs · Data validation |
Google / Coursera — 8 completed AI courses, October 2026.
View all eight Google courses
- AI Fundamentals
- AI for Brainstorming and Planning
- AI for Writing and Communicating
- AI for Research and Insights
- AI for Content Creation
- AI for Data Analysis
- AI for App Building
- AI for App Deployment
AWS Training & Certification — Foundations of Prompt Engineering, completed October 2026.