Technology Innovation Through Independent Execution
I founded three ventures spanning medical technology, autonomous systems, and financial services — each demonstrating the ability to identify market opportunities, execute complex technical solutions, and build sustainable business models.
Autonomous Underwater Vehicle Perception System
Project Overview
In December 2024, I launched a venture to develop a cutting-edge Autonomous Underwater Vehicle with dual-use applications in defense and civil markets, competing with high-end devices in quality while targeting at least 60% lower price point versus competitors.
Following initial system design, I strategically focused the MVP on developing the perception subsystem — a comprehensive multi-modal AI platform integrating 6 sensor types (camera, sonar, magnetometer, IMU, environmental, navigation) with advanced object detection pipeline featuring hierarchical classification across 5 categories and 49 object types. I executed complete technical development as solo founder while simultaneously managing strategic partnerships with German Maritime Authority (BSH), energy companies (Amprion, EnBW), and conducting comprehensive market validation across defense and infrastructure inspection sectors.
The outcome: a fully operational system achieving 95%+ safety-critical object detection with 2.8x performance optimization, complete synthetic data generation capability, and validated market interest—demonstrating that complex AI/ML systems can be successfully executed through strategic focus and technical excellence.
In July 2025, a strategic decision was made to pause the venture due to market timing and partner readiness challenges, while preserving the significant technical and business development achievements for future opportunities.
Technical Architecture of the perception subsystem
A Multi-Modal Sensor Fusion with Hierarchical Detection was designed with the following highlight
- 6 Sensor Integration: Camera, Sonar, Magnetometer, IMU, Environmental, Navigation
- AI/ML Pipeline: Progressive 3-stage training with safety-critical optimization
- Object Detection: 5 categories, 49 object types with hierarchical classification
- Performance: 2.8x optimization improvement (667ms → 239ms inference)
- Synthetic Data: 20,000+ realistic underwater images with effects simulation
Technical Achievements
🧠 AI/ML Implementation
- Complete PyTorch-based detection pipeline
- Cross-modal attention fusion mechanism
- Progressive training system with convergence optimization
🔬 Data Engineering
- Synthetic data generation with underwater physics simulation
- Multi-modal dataset management and versioning
- Realistic environmental effects modeling
⚙️ Systems Integration
- End-to-end pipeline from sensors to detection results
- Modular architecture with single responsibility design
- Comprehensive validation and testing framework
📊 Performance Optimization
- 95%+ recall for safety-critical objects (mines, UXO)
- Real-time inference capability with <200ms target
- Graceful degradation on sensor failure
Business Development
🏛️ Government Partnerships
- German Maritime Authority (BSH): Multiple department meetings
- Regulatory requirements analysis and compliance strategy
⚡ Energy Sector Engagement
- Amprion GmbH & EnBW: Technical demonstrations and interest validation
- Infrastructure inspection market sizing and value proposition
🎓 Academic Collaboration
- University partnership discussions for research collaboration
- University network development for technical validation
📈 Competitive Analysis
- Market positioning
- Technology differentiation and cost optimization strategy
Entrepreneurial Insights
The project demonstrated exceptional technical feasibility and market interest, but revealed the classic early-stage challenge: customers want proven products, while product development requires customer commitment.
Key Learnings:
- Solo technical execution is possible for complex AI systems
- Market timing and customer readiness are critical success factors
- Deep technical capabilities create significant competitive advantages
- Strategic pivoting preserves value while optimizing resource allocation
Strategic Outcome:
Transitioned focus to leveraging demonstrated capabilities in interim management and consulting roles, where technical depth creates immediate value.
Interested in the technical details?
The architecture shown here represents just the surface of a comprehensive AI/ML system. If you'd like to explore the complete tech stack, review detailed system architecture diagrams, understand the synthetic data generation methodology, or discuss how similar approaches could benefit your organization, let's connect.
I'm always happy to discuss the technical implementation with fellow engineers and technical leaders.
Trade Finance Innovation
Following extensive market research with suppliers and importers across multiple emerging markets, I launched Just-In Finance in April 2025 to address the €120B annual trade financing gap affecting African SMEs through innovative digital financing solutions.
I assembled and currently lead a cross-functional team of financial services, technology, compliance and market experts through early-stage development, while building strategic partnerships with banks, insurance providers, and trade organizations across target markets.
STATUS: Active development phase with established partnership pipeline and go-to-market strategy in progress.
