Engineering with a research mindset

I am a final-year Networks and Telecommunications Engineering student at INSAT, specializing in cybersecurity and working at the intersection of artificial intelligence, security, and network systems.

My main focus is AI engineering. I enjoy working across the entire lifecycle of an AI system: from model selection and dataset construction to fine-tuning, evaluation, failure analysis, optimization, and deployment. I have worked with LLMs and vision-language models using techniques such as LoRA and QLoRA.

During my internships, I worked on two very different applications of AI. I developed a fully local multilingual internal audit assistant using an open-source LLM, and later worked on vision-language models for real-time shoplifting detection from CCTV footage. These projects taught me that building a good AI system is not only about model performance, but also about understanding data, identifying failure modes, and designing reliable evaluation strategies.

Alongside AI engineering, my academic work explores the intersection between machine learning and cybersecurity. My main research project, FedGuard, investigates federated learning for real-time DDoS and intrusion detection, combining anomaly detection, distributed learning, network traffic analysis, and human-in-the-loop decision making.

What interests me most is the boundary between research and engineering: taking an idea from a paper or experiment, understanding its limitations, and turning it into something that can actually work in practice.