Hey There 👋

I'm Vladyslav Dykyi, an AI/ML Developer with expertise in:

• Development of AI models
• Deep Learning
• Audio Deep Learning
• Natural Language Processing (NLP)
• Computer Vision

C O N T A C T

• Email: vladdikiy17@gmail.com
• Website: https://portcode.at/
• Location: Linz, Upper Austria
• Phone Number: +43-688-640-595-08

P R O J E C T S

ADIFY AI

Founder of Adify, a Spotify playlist generator that creates playlists based on user prompts.

• Website: http://adify.pro/
• ADIFY AI uses the FAISS index and trained transformers, making it highly efficient. It has been trained on 10,000 playlists, allowing it to intelligently select songs and playlists from Spotify’s vast library based on user mood or activity. By entering a description of the desired playlist (e.g., "chill rock music for studying"), users receive personalized recommendations.

FLUXA AI

Founder of Fluxa AI, an advanced recipes generation website using state-of-the-art transformer models.

• Website: https://www.fluxa.pro/
• Fluxa is an innovative AI-powered platform designed to help users generate recipes based on the contents of their fridge. By simply uploading a photo of their fridge, Fluxa uses advanced image recognition models to detect the ingredients and suggest relevant recipes. It employs a custom-trained YOLO (You Only Look Once) model to analyze the image and identify ingredients, making it easy to create delicious meals based on what you already have.

WIZA TELEGRAM AI Assistant

Founder of the WIZA Telegram bot, a comprehensive travel assistant supporting 39 countries and 462 cities. - Features include top-rated restaurants and hotels, currency conversion, local taxi services, nearby hospitals, shopping centers, apartment rentals, emergency contacts, and an events section. - The bot utilizes a proprietary WIZA search engine and supports multiple languages for a seamless travel experience.

AI NYX Model

NYX is a Transformer-based neural architecture designed for autoregressive natural language generation tasks. The model leverages positional rotary embeddings, multi-head self-attention, and low-rank adaptation techniques for efficient, scalable training. The architecture is implemented using PyTorch and provides modular flexibility to suit a variety of NLP applications, including text generation, completion, and language modeling.

S K I L L S

• AI/ML Model Development
• Deep Learning (CNNs, RNNs)
• NLP and Audio Processing
• Computer Vision
• Model Optimization and Deployment
• Data Analysis and Preprocessing
• LLMs and Chatbot Development
• Web-Scraping and Automation
• Python, TensorFlow, PyTorch