Welcome! This account is dedicated to my personal projects, separate from my work-related activities. This helps me keep an accurate track of my personal project progress and maintain a separate daily contribution streak, distinct from my work contributions.

For work-related projects, visit my work GitHub account: aapotapov

I'm currently working on several innovative and open-source projects, and I warmly welcome collaborations! Feel free to reach out if you're interested in contributing to any of the projects below:

1. PhriniFluent.com: an open-source service for language learning/words memorization powered by OpenAI. Also serves as my full stack web development demo.

   • PhriniFluent Telegram Bot - PhriniFluent's experience as a Telegram bot
   • PhriniFluent Backend - PhriniFluent's Backend written in Django
   • PhriniFluent Frontend - PhriniFluent's frontend written in Angular
   • QuizBot Telegram Bot - Finished and ready to use first iteration of PhriniFluent as a Telegram bot

2. ExcelTableCNN: Open-source library for automatic table detection in Excel sheets based on Computer-Vision Deep Learning techniques.

3. ProjectTextSuite: An open-source library designed for QnA LLM pipeline over a projects database folder, supporting multiple file types like word, pdf, pptx, etc with file version control and ability to do automatic question-answering pipelines over large table files like xlsx. Suite consists of the following 6 packages all of which could be used independendly for your purposes (i'm pretty happy with it's modular design):

   • TextTableScoop - file-to-text and table-to-csv parser that safely parses the files with the use of LibreOffice in order to extract text and tables in csv format from the files.
   • VecMetaQ - server wrapper over a FAISS vector database with encapsulated embeddings and metadata storage in order to achieve the fast similarity search and ease of use at the same time.
   • ProjectTextAgent - file observer of you project directory as a Docker agent written in Go that updates the database for current state of your files.
   • xlsx2pandas (PandasXcelerator) - xlsx file parser in order to smartly obtain the data in a flat form in order to import it into a pandas dataframe.
   • RelaLLM (Coming soon!) - Multi-Header Dataframe Processing with LLM-Powered Relational Database Mining.
   • ProjectTextQnA (Coming soon!) - question-answering interface over your project data in order to perform LLM qa and text-to-sql pipelines. Includes support of both self-hosted LLMs and high quality ones such as OpenAI API.

4. My personal blog (In progress, coming soon!): My safe place to post hot takes on modern development, IT trends, entrepreneurship.

   • FlagroTown Blog Backend - authentication (oauth2), blog management (PostgreSQL, AWS S3) and analytics services implemented in Flask
   • FlagroTown Blog Frontend - blog's static frontend written in Svelte with SvelteKit
   • FlagroTown Blog Repository - new articles are handled by upload to a git repository and blog management is activated via github webhooks. Also this allows me to host my blog simultaneously in two places :)

5. SummaryForEveryone (Currently on pause): An open-source library for easy and convenient summarization of text, audio, and video using local LLM models.

6. treefeeder: My take on easy communications with your LLM/AI assistant over your codebase.

Also please check out my github gists!

I am open to collaborations on any of the projects above. Whether you want to contribute to the code, discuss ideas, or offer suggestions, feel free to open an issue or send a pull request!

I'm a Data Analyst, LLM Engineer, and Developer specializing in Python and C++. With 3 years of hands-on experience in ML and backend development, I've been contributing to projects supported by brands like Huawei and refined my skills in backend development, focusing on innovative tech solutions. Currently, I'm enhancing the realms of Large Language Models and Natural Language Processing.

Since September 24th 2023, I have embarked on a focused journey towards implementing my vision for open source. Each single day since then is marked with a commitment to contribute, and as of now, I'm proud to maintain a streak of more than a year of every day contributions. This consistent effort reflects my dedication to implement my ideas and goals in the open-source community. Every commit, issue, and pull request is a step towards making a meaningful impact and bringing my vision to life.

📍 Location: Podgorica, Montenegro

• Backend Development: C++, C, Golang, Python, Django, Django Rest Framework (DRF), FastAPI, AWS
• Frontend Development: Angular, Svelte, SvelteKit
• Data Engineering: Pandas, PostgreSQL, MongoDB, Motor, Openpyxl, win32com, SQLite, SQLAlchemy
• CI/CD: Git, GitHub Actions, Docker, Kubernetes
• LLMs/NLP Tasks: Transformers, Bert, GPT, T5, Llama, Pytorch, HuggingFace, langchain, NLTK, Spacy
• Parallel/Distributed Computing: PThreads, OpenMP, MPI, CUDA, RabbitMQ

• Bachelor's degree in Applied Mathematics and Computer Science, Lomonosov Moscow State University (2019 - 2023)

• Role: Data Analyst, LLM Engineer
• Contributions: Developed automation tools, integrated advanced LLM technologies, developed MVPs for machine learning products, and proposed innovative solutions to reduce human hours and company costs.

• Role: C++ Developer
• Contributions: Developed optimized C/C++ backend implementation for the GraphBLAS standard and implemented a series of optimizations for matrix-vector multiplication.

• Music production, playing guitar, bass, drums, and having fun with MIDI

• 🌐 My personal website: flagrotown.com
• 🌟 LinkedIn: linkedin.com/in/flagrogrey
• 📸 Instagram: @flagrogrey