Hi!

My name is Atanas Tonchev and this is my personal website. Here you can find out about my experience, projects, research, and general interests. The Research and Software pages contain more detailed information about my experience and foci, while also pointing you to relevant posts in my blog.

I am interested in the intersection of social research and data science. I am the co-founder and data consultant at Polus Research, a consultancy that helps small and mid-sized businesses integrate data science, data engineering, and machine learning into their operations.

I have also been working on several personal projects at the intersection of AI, NLP, and software development. You can read more about them in the Projects sections of the Software and Research pages.

Research

I have an academic background in social research, with a focus on both qualitative and quantitative methods. I have conducted research on topics such as the impact of religious identity on political polarization, the risks and potentials of generative AI for humanities education, and leveraging graph analysis as a tool for monitoring online disinformation.

I have been following the development of large language models (LLMs) since a few months before ChatGPT was unveiled in November 2023, and I have been actively following the fields of LLM interpretability and AI social impact. I am particularly interested in how these models can be used to improve our understanding of human society through language itself.

More recently, I have been researching the intersection of Orthodox Christian online content and far-right metapolitics. My project, "The Most Masculine Christianity: Metapolitics and Far-Right Normalization in Orthodox Christian Online Content," combines NLP, vector embeddings, and graph analysis with qualitative social research to map how Orthodox online spaces become vectors for radical political normalization. This research has produced a working proof-of-concept dashboard, Orthosphere, available at orthosphere.ai-tonchev.com.

Software

I believe the most important skill for a software engineer nowadays is the ability to identify and learn the proper tools for a job. This process of research is something I have enjoyed in all my previous projects.

I got into software development by facing new problems and finding solutions to them. I began by using Python to analyse quantitative data, then I gradually had to familiarize myself with data modelling and warehousing, big data cloud computing, advanced visualization and automated infographics, and, eventually, web development for the website you are scrolling through.

Soft Skills

I have a keen sense of ownership on any project I work on. This means I have always been concerned with the impact my work has on all its stakeholders: users, business owners, and society at large. I have not only sought to create software infrastructure that works, but also ensure I leave behind myself a process that is sustainable and maintainable, so that the business can continue to operate effectively after I am gone.

I have a strong interest in how data science and AI can be used to improve business processes, enhance decision-making, and drive innovation. With Polus Research, I have been particularly interested in how these technologies can be used to create value for small and mid-sized businesses, which often lack the resources to fully leverage the potentials of data science.

Last but not least, I keep up with the latest research and developments in the field of data science and NLP in particular, while constantly expanding my knowledge of various fields, including advanced statistics, web development and network architecture, UI design etc.

Music

I am also an amateur musician and producer with a passion for Bulgarian folk instruments, electronic music, jazz clarinet, and jazz more broadly. I am proud to be a member of the Tenebris project. I hope to soon expand my website with a section dedicated to my music, where I will share some of my tracks and projects.