Point Cloud Processing Expert in Germany

Development of a route planner for incident communication.

• Development of the REST API
• Set up a patch system for maintaining the routing graph
• Expansion of the testing infrastructure
• Performance and memory optimization (JMeter, JFR)

Technologies: Java 21, Spring Boot, JGraphT, Flyway, MapStruct, Caffeine, ShedLock, JMeter, Kubernetes, JFR

• Built a Trend-Scouting AI system to automate technology intelligence in power electronics and semiconductors, combining Azure OpenAI with LangChain, Scrapy-based web crawling for structured, noise-free data acquisition, and automated PDF reporting for internal R&D use. Developed a FastAPI-based (Uvicorn) web application to validate LLM outputs, test prompt strategies, and enable interactive system evaluation.
• Developed a real-time STM32 binary telemetry debugger with a PyQt-based GUI, featuring header-based frame synchronization, anomaly detection, template-driven payload decoding, time-aligned buffering, and live signal visualization.
• Developed an AI-driven power inductor designer using surrogate regression models for accurate electromagnetic and thermal prediction. Integrated multi-objective NSGA-II optimization to generate efficient, manufacturable designs.

• Developed a CNN-based pedestrian detection system using LiDAR data
• Created Python scripts for bounding boxes, dataset labeling, and data conversion
• Evaluated model performance on datasets with point clouds

• Developing methods for monocular 3D facial reconstruction and personalized geometric modelling from mobile imagery
• Building learning-based approaches for facial shape estimation, video-based facial analysis, and privacy-preserving visual learning

• Developed point cloud classification and segmentation models for dental applications.
• Designed domain adaptation techniques that improved F1 score by 0.1 on a new clinical domain.
• Worked with 3D geometric data and production-scale ML pipelines.

• Independently initiated and deployed automated validation workflows using Python, cutting manual processing by 58% and improving efficiency
• Developed a machine learning model for synthetic defect generation, reducing downtime and production costs; deployed locally and via Databricks and Azure AI Factory
• Utilized a small dataset of image data from the production lines and extended this dataset with training on models like cycleGAN and pix2pix
• Built and optimized the Linux-based development environment for training 3D models; maintained reproducibility via GitHub
• Presented technical insights to cross-functional teams (engineers, QA, project managers), ensuring alignment of ML solutions with operational needs

• Introduced an integrated approach for structural damage detection across concrete, steel, and glass using advanced technologies such as LiDAR and thermal imaging. Highlighted cross-material interactions to enhance diagnostics and enable predictive maintenance.

• Developed an NLP-based medical note simplifier that transforms complex clinical instructions into plain, child-level English. Applied prompt engineering with Flan-T5 transformer models to extract patient-relevant actions and rephrase them into clear to-do items. Built dual Flask and Tornado backends with a printable web interface.

• Developed an evaluation framework for AI-generated multiview driving videos intended for perception and embodied-AI/VLA-related training workflows.
• Designed automated checks for temporal coherence, cross-camera consistency, semantic correctness, and multiview geometric quality, exposing failure modes relevant to autonomous systems.
• Combined classical computer vision, learned visual representations, and vision-language models to convert complex video artifacts into measurable engineering signals.
• Built repeatable benchmarking and failure-analysis workflows to support model comparison, data-quality decisions, and system-improvement discussions.

• Developed a Python-based synthetic data generation pipeline in Blender to simulate complex human-forklift interactions for robotic perception and AI model training.
• Designed and modeled 3D industrial digital twins to support depth estimation, stereo vision, and safety analysis workflows.
• Collected and processed LiDAR, laser, and photogrammetry point clouds to generate accurate 3D maps for environment reconstruction and ground-truth data creation.
• Developed and deployed YOLOv8-based pose estimation and depth perception algorithms using PyTorch and OpenCV, optimized for GPU clusters and NVIDIA Jetson platforms.
• Integrated and validated AI modules in ROS-based robotic environments, ensuring real-time performance and interoperability.

• Topic: Object Detection and Semantic Segmentation for (AUV) Systems using Transformer-Based Vision Models and Sensor Fusion.
• Designed and implemented an end-to-end multi-sensor fusion perception pipeline (Camera, LiDAR, IMU) in ROS
• Developed CNN-based Machine Learning model (YOLOv8) and Transformer-based vision models for real-time object detection
• Processed and clustered 3D LiDAR point clouds using DBSCAN, RANSAC, and voxel grid filtering to enable robust object localisation in noisy environments.
• Designed Bayesian Network models (GeNle) for probabilistic reasoning and sensor-level decision fusion under uncertainty.
• Applied Kalman filtering for sensor state estimation, temporal alignment, and smooth object tracking, reducing false positives in safety-critical scenarios.
• Evaluated system performance under realistic driving dynamics, improving tracking stability and overall perception robustness.
• Built deep learning pipelines for training, validation, and performance evaluation of perception models using sensor data.

• Vram.run: Rust, TypeScript, HF Inference API with 19 providers, 220+ HW configs, and 30+ cloud GPUs. Search a model to see which API providers serve it, which GPUs can run it locally (and how fast), and what cloud rental would cost. Or search your hardware and see what fits. Also includes a Rust CLI.
• Psychotron: JavaScript, Web Audio API, AudioWorklet, Canvas 2D. Front-end for flash fiction audiobook with Web Audio DSP chain featuring pitch-shifting, 12-voice chorus, flanger, 13-band EQ, and convolver reverb. Includes a 2D canvas effect morphing engine and synchronized teleprompter.
• RecentWork: Swift, macOS, FSEvents, launchd. macOS daemon that watches project directories and maintains a flat folder of symlinks to recently modified files. Homebrew installable.
• Mini-llm: Bash, macOS, launchd, Ollama, llama.cpp, MLX, Open WebUI. Single command that turns a Mac Mini into a headless AI server.
• ThatSlop: JavaScript. Chrome/Firefox extension for AI content detection on LinkedIn and Twitter.
• Smux: Bash, tmux. Human-friendly tmux wrapper that is Homebrew installable.
• Learn Rust Course: Rust. Course on Rust’s memory model for C++ programmers, written from experience of transitioning from C++ to Rust at Irreducible.

• Serve as AI lead for life-science solutions, integrating advanced AI models directly into company workflows and ensuring seamless deployment.
• Design and implement deep learning architectures (PyTorch, Keras) for complex biomedical challenges, including cell segmentation, multimodal omics analysis, and prediction of point clouds.
• Develop and deploy robust LLM-based systems, including RAG architectures and agentic workflows using LangGraph, to facilitate natural-language interaction with complex medical data.
• Lead cross-functional initiatives to apply foundation models and explainable AI (xAI) to clinical and evolutionary algorithms.

• Planning and design of plants using AutoCAD Plant 3D, Inventor, Advance Steel, Navisworks
• Review of planning documents such as concepts and design reports, R&ID
• Processing and creation of pipe classes
• Creation and development of piping routes in 3D plant models
• Creation of isometrics of piping systems and special supports
• Coordination of involved trade specialists, foundation planning
• Consulting in project development and execution of projects in plant engineering, chemical engineering, and refinery construction
• Coordination with suppliers for basic engineering and tender preparation
• Management of all relevant engineering activities to create a technical concept (scope definition)

• Implemented stereo camera calibration and applied incremental Structure from Motion (SfM) algorithms to build a detailed 3D model of a forklift for an AR-enabled Smart Forklift project
• Developed OpenCV-based preprocessing that improved data accuracy by 20% and enhanced the analysis of industrial videos