Software Engineer · MSc Data Science Candidate · Creator of Denxhinjo Labs
Building AI-powered applications, distributed systems, high-performance software, and privacy-aware machine-learning solutions.
Portfolio · LinkedIn · Repositories
I am a software engineer and MSc Data Science candidate at the University of Basel.
My work combines software engineering, artificial intelligence, distributed systems, privacy-preserving machine learning, and high-performance computing. I enjoy transforming technical ideas into practical systems—from inventory platforms and AI support tools to differentially private models and parallel algorithms.
Through Denxhinjo Labs, I document selected products, academic research, and engineering experiments.
- MSc Data Science candidate at the University of Basel
- Interested in AI engineering, distributed systems, HPC, and full-stack development
- Currently exploring peer-to-peer navigation and traffic dissemination
- Based between Basel, Switzerland and Tirana, Albania
| Project | Description | Main Technologies |
|---|---|---|
| FAB Construction IMS | Inventory and operations platform supporting construction workflows, access control, stock tracking, locations, and work-process management. | FastAPI, React, TypeScript, PostgreSQL, Docker |
| High-Performance Computing | Parallel implementations and benchmark analysis across shared-memory, distributed-memory, and GPU programming models. | C++, CUDA, MPI, OpenMP, OpenACC, Slurm |
| Privacy-Preserving Credit Prediction | Evaluation of differential-privacy approaches through model utility, membership-inference resistance, and fairness analysis. | Python, PyTorch, Opacus, Streamlit |
| Adaptive Optimization | First-order and quasi-Newton optimization algorithms implemented and benchmarked from scratch. | Python, NumPy, Streamlit, Docker |
| RYCO AI Support Demo | Public demonstration of an AI-assisted support workflow with conversational assistance and ticket automation. | JavaScript, Groq API, Supabase, Resend, Vercel |
| Geo-P2P Routing Build | Automated OpenTripPlanner graph-building workflow using public-transport and map data for the Basel region. | GitHub Actions, Docker, OpenTripPlanner, GTFS, OSM |
I am exploring a decentralized navigation system in which nearby users exchange traffic conditions and hazard alerts directly rather than relying entirely on a centralized traffic-data service.
Current areas of investigation include:
- Rust and iroh networking
- Geographic topic-based dissemination
- Peer discovery and NAT traversal
- Location privacy
- OpenTripPlanner routing
- GTFS and OpenStreetMap integration
I am also interested in the practical relationship between:
- Differential privacy
- Model utility
- Membership-inference resistance
- Fairness
- Deployment constraints
Programming
Python · TypeScript · JavaScript · Rust · C++ · Java · PHP · SQL
AI and Data
PyTorch · Opacus · scikit-learn · Pandas · NumPy · Jupyter · Streamlit
Distributed Systems and HPC
MPI · OpenMP · CUDA · Slurm · AWS · Neo4j · Apache Spark
Web and Backend
React · Next.js · FastAPI · Laravel · Spring Boot · ASP.NET Core
Infrastructure
Docker · GitHub Actions · Linux · Vercel · PostgreSQL · Supabase
Detailed case studies, live demonstrations, research projects, and engineering experiments are available at:
For professional opportunities, technical collaboration, or research discussions: