Portfolio 2026
Martin Zhelyazkov
I build AI-powered systems with a security-aware mindset, connecting research, backend engineering, and practical product workflows.
AI Research
LLM benchmarking and RAG
Backend
APIs, data flow, integration
Security
Reliable, careful systems
Current Focus
Research-backed AI with practical delivery
Internship Benchmark
Compared LLaMEA, LLM4AD, and frontEASE with one reproducible setup.
AI Backend Work
Grounded models, training pipelines, and backend integration for team projects.
Cyber Security Direction
Building toward AI systems that are useful, robust, and security-conscious.
About Me
Introduction
Hi, I'm Martin Zhelyazkov, an AI developer with a passion for Cyber Security, focused on building robust, scalable, and efficient systems that power modern web applications.
Why Applied Computer Science, AI, and Cybersecurity
I decided to study applied computer science because I have always been fascinated by technology and the inner workings of things. As for cybersecurity, I believe that in today's digital age, understanding security is crucial for protecting data and systems from threats. I want to be at the forefront of this field, ensuring safety and privacy in an increasingly connected world.
Unique Hobbies
Outside of development, I enjoy fitness and sports, particularly calisthenics, as well as learning new languages and skills that involve creative thinking and hand-eye coordination. I am especially interested in card tricks and archery, which require precision and focus. Finally, I enjoy learning new things and experimenting with new technologies.
Future Dreams & Professional Ambitions
I aim to become a lead AI developer, designing high-performance systems and APIs for global-scale applications. I also plan to contribute to open-source infrastructure tools and help grow the next generation of AI developers through mentoring and collaboration.
Internship Highlight
Benchmarking LLM-Driven Algorithm Design
During my internship at Tomas Bata University in Zlin, I worked as an AI developer and research intern on a benchmark for automated algorithm discovery.
I built a repeatable Python evaluation pipeline for LLaMEA, LLM4AD, and frontEASE, connected the frameworks to the same local LLM setup, validated generated algorithms, and turned the results into figures and documentation.
Internship Output
Benchmark Harness
Shared tasks, seeds, timeouts, adapters, raw JSON results, and resume support.
Result Analysis
Notebook-generated figures and comparisons across the frameworks.
Portfolio Documents
Project planning, realization document, and reflection are available on the internship page.
Featured Projects & Activities
A focused look at my internship, machine learning coursework, and international BIP work.
AI Research Internship
LLM Algorithm Benchmark
Internship Research Project
A reproducible benchmark comparing LLaMEA, LLM4AD, and frontEASE for automated algorithm discovery.
View InternshipMachine Learning Course
ML, NLP & Cloud
Machine Learning Projects
Three course projects where I worked as project leader across computer vision, NLP, and cloud-based machine learning.
View ProjectsBIP Beja, Portugal
Cheese Tracking
Smart Cheese Tracking System
An international BIP project where I gathered data and developed an AI model to track cheese ripening.
View ProjectCurriculum Vitae
Download my CV to learn more about my skills, experience, and qualifications. Feel free to reach out for collaborations and opportunities.
Download CV