• Building an AI unsupervised algorithm for Museum object detection, segmentation and identification an Search engine for Museum objects
• Advising for the AI Architecture of the system
• Building an online and offline Search algorithm for specific museums

• Advising the dataset acquisition
• Build the AI infrastructure to develop the detection and counting problem in urban scenes
• Building the AI architecture for crowd-tracking with multiple cameras

• Lead the development of state-of-the-art computer vision algorithms to detect fires both at night and during the day
• Led the research and development of multi-frame detectors
• Led the research and development of multi-frame from a static single-camera distance estimation
• Mathematical modeling of fire initiation and growth.
• The role was focused on leading the detection transformation. The 20-month journey reduced several million in costs and achieved real-time smoke and fire detection.
  

• Integrated tfRecord reading component for multi node training
• Integrated MlFlow visualization
• Integrated Azure ML MLflow
• Training using Ray.io

• Mentor and teach student
• Advisor for AI projects in the fintech and manufacturing sector
• Providing training to banks, financial institutions, and manufacturing factories in Artificial Intelligence adaptation
• Successful industry/academia students projects:
  • Integrate a Generative AI solution at a large bank
  • Integrate a Computer vision solution in Iron classification and real-time measurements in a large iron manufacturing company
  • OCR for a stock Iron stock management from the excavators.
  • ?Information extraction from Financial reports
    

• Constructed an algorithm for false Artwork detection
• Constructed a fast CPU feature extractor and matcher to identify non-original artworks that were completed in less than 2 seconds.
• Constructed an algorithm for classifying artworks in several categories and the quality of the artwork.
• Developed Light glue algorithm and statistical feature extraction to analyze the artworks in real-time.
  

• Constructed OCR and Intelligent Document processing pipeline for collectibles (trading cards)
• Developed several object detectors to aid document classification and information processing.
• Extracted and structured the information of several non-standard card templates
• Assisted in creating the AWS infrastructure to deploy the IDP processes

• Constructed computer vision algorithms for object detection in electromagnetic wave-generated images using Ground Penetrating Radar.
• Conducted experiments and built algorithms for simulated and real data.
• Guided stakeholders in acquiring real-world data for algorithm development.

• Led a team of five people, (3 annotators, 1 junior CV engineer, and a backend engineer)
• Developed computer vision algorithms for object detection, counting, and tracking in warehouse settings.
• Achieved 3D reconstruction of objects using multiple images and graph theory.
• Built a graph model of the objects to improve the robustness of the object detector
• Guided and mentored three annotators and one junior Computer vision engineer

• Created image search/matching engine for sports and game cards, utilizing contrastive and triplet loss metrics.
• Used contrastive and triplet loss metrics to build an image embedding system.
• Build a hard example mining system to improve the algorithms
• Implemented FAISS indexer for efficient image indexing.

• Developed an unsupervised learning algorithm for defect detection of copper plates by creating a metric learning loss function.
• Constructed a computer vision algorithm for detecting the place where two parts should be joined.
• Constructed computer vision algorithms for object detection and tracking in X-ray images of manufactured parts.
• Optimized detection to process over ~17 images per second.
• Communicated with clients to ensure high-quality data and annotations.

• Built an OCR automatic parsing system for data uploaded to Google Drive.
• Managed various templates and leveraged cloud-based tools.

• Developed computer vision algorithms for card and coin defect evaluation, collaborating with operations and developers.
• Designed and managed the end-to-end process, encompassing data acquisition, scanning, annotation, and the complete computer vision algorithm development lifecycle through to deployment.
• Ensured the seamless functionality of computer vision algorithms by actively communicating and coordinating efforts with various departments.

• Developed object detection and tracking algorithms for 3D Computed Tomography images. Reduced the human (worker) time from ~120 minutes to 5 minutes.
• Created invoice data extraction algorithm for German digital post service.
• Constructed NLP algorithms for research paper understanding and computing their correlation to UN sustainable development goals
• Proposed ML/Data Science projects for clients.

Built an invoice data extraction algorithm for a private German digital post service

Built an object detection and tracking algorithm for counting objects in 3D Computed
Tomography images of vehicle parts. Reduced the human (worker) time from ~120 minutes to
5 minutes

• Built a high-accuracy pore detection algorithm for 3D CT-Scans with 99.9% accuracy.
• Designed change detection system in airplane cabins using RF signals.

• Developed state-of-the-art (for small datasets) Semantic Scene Segmentation algorithm for façade images
• Created dataset for detailed facade segmentation and 3D building models.
• Built pipeline for façade segmentation from data acquisition up to deployment.

• Improved (by more than 50%) matching algorithm accuracy for user profiles and problem descriptions.

Taught Courses: Calculus 1 and 3, Discrete mathematics 1 and 2 and Basics of IT