Motivated Computer Science student (3/4 year) with a strong foundation in Python, C++, and AI Systems. Passionate about DevOps practices and the operationalization of machine learning systems, from designing automated CI/CD pipelines and containerized cloud deployments to building and benchmarking local LLM inference architectures. I combine academic rigor with hands-on experience shipping end-to-end solutions across the full development lifecycle..
B.Sc. in Computer Science | Faculty of Mathematics and Computer Science
Focus: Data Structures, Algorithms, OOP, Operating Systems, Artificial Intelligence, Completed CCNA 1 & 2 curriculum
Built a hybrid AI pipeline combining a local DeepSeek R1 8B model with symbolic computation (SymPy) and RAG (ChromaDB) to solve mathematical problems. The edge-based system achieved 100% accuracy on GSM8K and 90.3% on MATH500, matching cloud-based Gemini 2.5 Flash.
Tech: Python, Ollama, SymPy, ChromaDB, RAG, Flask, LLM Orchestration, Multi-node HTTP Comunication
Designed and shipped a production-style CI/CD pipeline for a containerized .NET 10 + Angular 19 application built on Clean Architecture. Implemented multi-stage Docker builds, a self-hosted GitHub Actions runner publishing images to a private Nexus Docker Registry, and automated SSH-based deployment to a Linux web server via docker compose. Separated dev/prod environments through Docker Compose overrides and centralized structured logging with Serilog.
Tech: Docker, Docker Compose, GitHub Actions (self-hosted), Nexus Repository, Nginx, .NET 10, ASP.NET Core, Angular 19, PostgreSQL, Linux/SSH
Developed a progressive portfolio of embedded systems projects using C++ and Arduino, ranging from analog signal processing and PWM control to complex applications involving Finite State Machines, hardware interrupts, and sensor integration.
Tech: C++, Arduino Uno, Circuit Design, Interrupts, I2C/SPI/Serial Communication
Engineered a cybersecurity tool that converts raw hexadecimal .bytes files into image representations to detect malware patterns. Trained a CNN achieving ~92% accuracy.
Tech: Python, TensorFlow, Keras, Scikit-learn, Image Processing
Built a full-stack automated analysis tool that aggregates financial news. Integrated the DeepSeek R1 AI model hosted on Azure to perform sentiment analysis.
Tech: Python, Flask, Azure Cloud, PostgreSQL, LLM Integration
Designed a desktop application for logistics management using the QT Framework. The project emphasizes strict OOP principles.
Tech: C++, QT Framework, GUI Design
Authors: Bogdan-Marius Dinu, Ciprian-Mihai Ceausescu, Cristian-Hacic Kevorchian
Accepted - 30th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES 2026)
ALGORITHM Centru de Meditații și Dezvoltare Personală
MAGIC HEATING SYSTEM SRL
NEW DIAMOND ROAD SERVICE SRL
FORTES