Kamen Zhekov

Kamen Zhekov

Software and AI Engineer

Contact Me

About Me

Curious, driven and with a handy Master's Degree in Applied Computer Science from the Vrije Universiteit Brussel. I have 2 years of Machine Learning Engineering experience acquired through my work as a consultant for ING through Exellys and my work as an AI Engineer at Contract.Fit. I currently work at Odoo as a Python & JS Software Engineer. Throughout my career, I have designed and integrated various features in a system for automatically processing documents, extracting information and classifying them using regular algorithms as well as Machine Learning, and I've migrated several Machine Learning and Data Science models to Google Cloud using the various technologies available for data processing, model execution, automated deployment and scheduling. During my studies, I've worked on projects for desktop applications using Python, Java, C++, C and ASM that implemented common modern programming practices such as Test Driven Development, agile methodology and version control. I also have experience in general dev-ops tasks, machine learning and web development. My Master's thesis consists in designing and implementing a complex document classification system using handcrafted and automatic feature extraction from images and text to label documents, demonstrating state-of-the-art performance for a well-explored application.

Work Experience

Python & JS Software Engineer - Odoo (2021 - Present)

Worked on the open-source online web-based website and e-commerce builder, using various JS and Python technologies to add new functionalities as well as improve and optimize existing ones.

Machine Learning Engineer - ING through Exellys (2020 - 2021)

Worked on migrating Machine Learning and Data Science models to Google Cloud, using the various technologies available for data processing, model execution, automated deployment and scheduling.

AI Software Engineer - Contract.Fit (2019 - 2020)

Worked on integrating a machine-learning based document classification system, improving confidence levels and automation for existing predictions, document splitting and various dev-ops tasks.

Master's thesis - VUB (2019 - 2020)

Designed and implemented a state-of-the-art machine-learning based document classification system using handcrafted and automatic image and text feature extraction.

Software Engineer Intern - Contract.Fit (2019)

Worked on improving an existing product for automating document processing (classification and information extraction) by implementing algorithms for finding and lemmatizing separable verbs in dutch using NLP, dynamic thresholding in multi-class multi-label prediction models, tax details extraction from invoices and receipts, etc.

Latest Projects


GPT3 Horoscopes

GPT3 Horoscope Generation

Training and deploying GPT3 (Neo) for horoscope generation using scraped horoscopes, creating an API, Dockerizing and deploying container to Google Cloud Run.

Read blog postWebsite & Frontend CodeBackend Code

EDSR Implementation

Super Resolution EDSR Implementation

Deep-learning based algorithm that performs super-resolution upscaling (x2, x3 or x4) on any image, and corrects pixel irregularities with an algorithm based on bilinear interpolation.

Check it out

Microcalcification-Analysis-SVM

Machine Learning-based Breast Cancer Detection

Machine-learning assisted algorithm that helps detect whether a patient has breast cancer or not, based on the analysis of characteristics extracted from the 3D models of microcalcifications present in the patient’s breast tissue.

Check it out

Spyware Project

Spyware Project

Spyware that works by mimicking a desktop PC optimization app, running a server in the background which sends sound recorded from the default microphone to any machine that connects to it. It sends the infected machine's IP address to a gmail inbox.

Check it out