Generative AI · Machine Learning · Agentic AI · Computer Vision · Applied Data Science
I am an M.Sc. Data Science student at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) with a background in Computer Science and a growing focus on Generative AI, machine learning, agentic systems, computer vision, and applied AI research.
My projects span multimodal generative AI for education, multi-agent LLM systems, deep learning, computer vision, geospatial data analysis, and public-health analytics.
Alongside my studies, I work in User Acceptance Testing at Siemens Healthineers and contribute to digital and IT initiatives at START Nürnberg.
I am particularly interested in moving from building AI applications toward studying questions around:
- LLM reliability and evaluation
- Agentic AI systems
- Multimodal AI
- Human-AI interaction
- Trustworthy and responsible AI
- AI for education and healthcare
- Applied machine learning
- Scientific and data-driven AI systems
Bachelor's thesis project developed at the University of Pécs.
A university-focused AI platform integrating multiple generative modalities into a single interface.
Features
- Conversational AI
- Code generation
- Image generation
- Video generation
- Music generation
- University-authenticated access
Technologies:
Next.js · TypeScript · OpenAI API · Replicate · Clerk · Tailwind CSS
An experimental multi-agent AI system for discovering, analysing, validating, and monitoring regulatory information.
The system separates tasks across specialised agents for information discovery, regulatory matching, planning, validation, and monitoring.
Technologies:
Python · LangGraph · Gemini API · LangChain · ChromaDB · LlamaIndex · Sentence Transformers · Streamlit
🔗 ComplianceNavigator Repository
Computer-vision system for detecting road cracks from drone imagery and mapping detected damage geographically.
Features
- U-Net image segmentation
- ResNet34 encoder
- Road-crack detection
- GPS extraction from EXIF metadata
- Interactive spatial mapping
- Full-stack inference interface
Technologies:
Python · PyTorch · U-Net · ResNet34 · OpenCV · FastAPI · React · Folium
🔗 Drone Road Damage Repository
A spatial-data quality assurance pipeline for road-network datasets.
Features
- Geometry validation
- Dangling-edge detection
- Disconnected-component analysis
- Spatial intersection checks
- CRS transformations
- GeoJSON output
- Interactive maps
Technologies:
Python · GeoPandas · FastAPI · Folium · GeoJSON · Docker
Neural-network experiment studying the effects of regularisation on a noisy three-class spiral dataset.
Compared baseline and regularised neural networks using:
- L2 regularisation
- Dropout
- Early stopping
- Learning-rate experimentation
- Decision-boundary analysis
Reported test accuracy improved from 91.7% to 96.67% after regularisation.
Technologies:
TensorFlow · Keras · scikit-learn · NumPy · Matplotlib
🔗 Spiral Classification Repository
Applied data-science project combining global child-mortality, malnutrition, demographic, and regional datasets.
Methods
- Data cleaning and integration
- Exploratory data analysis
- Correlation analysis
- Clustering
- Regression
- Geographic and temporal comparisons
Technologies:
Python · Pandas · NumPy · scikit-learn · Matplotlib
Computer-vision pipeline for analysing football videos.
Includes:
- YOLO object detection
- Player and referee detection
- Ball tracking
- Team classification
- Camera-motion estimation
- Perspective transformation
- Speed and distance estimation
- Clustering and tracking
Technologies:
Python · YOLO · PyTorch · OpenCV · scikit-learn · Roboflow
🔗 Football Analysis Repository
I am currently interested in exploring research involving:
Generative AI
LLMs · multimodal models · model evaluation · information fidelity · AI-generated educational content
Agentic AI
Multi-agent systems · tool-using agents · retrieval systems · autonomous workflows · agent evaluation
Trustworthy AI
Reliability · robustness · interpretability · evaluation · responsible AI
Machine Learning & Deep Learning
Neural networks · regularisation · computer vision · representation learning
Applied AI
Education · healthcare · scientific applications · geospatial analysis · public-interest data science
M.Sc. Data Science
Friedrich-Alexander-Universität Erlangen-Nürnberg
2025 – Present
B.Sc. Computer Science
University of Pécs
2021 – 2024 · GPA: 4.48 / 5.00
Working Student — UAT Support
Siemens Healthineers · Erlangen, Germany
Managing Director — IT Department
START Nürnberg · Nürnberg, Germany
Front-End Developer Intern
University of Pécs — STUDEV · Pécs, Hungary
Stipendium Hungaricum Scholarship
Awarded for my Bachelor of Science studies at the University of Pécs.
- 📧 Email: abbasiussayedshakeel@gmail.com
- 💼 LinkedIn: linkedin.com/in/ussayed
- 💻 GitHub: github.com/abbasi303
