Данило

DANYLO KHOMYSHYN
ML ENGINEER

PROFILE
ML Engineering student at Lviv Polytechnic National University (Artificial Intelligence
Systems). I build backend systems and AI pipelines — from RAG and LLM integrations to
REST APIs and vector databases. I focus on writing clean, production-ready code and want
to grow inside a strong engineering team.

CONTACT
WORK EXPERIENCE
[відкрити контакти](див. вище в блоці «контактна інформація»)
[відкрити контакти](див. вище в блоці «контактна інформація») ML / AI Engineer — NDA
• Contributed to the development of a production agentic AI system for automated
github.com/khomyshyn13 promotional video generation, orchestrating multiple AI-driven stages within an end-to-end
workflow.
telegram.org/khomyshyn13 • Designed and implemented components of the agent-based pipeline for coordinating
generation tasks, intermediate outputs, and multi-step processing.
• Integrated AI models and external services into a unified production workflow with a focus
SKILLS on reliability, modularity, and maintainability.
• Developed backend components and APIs supporting AI inference and asynchronous
Python (OOP, algorithms, data processing.
structures) • Collaborated with the engineering team on bringing AI prototypes into production-ready
RAG pipelines, LLM integration systems.
Project-specific architecture, models, and implementation details are protected under NDA.
PyTorch, TensorFlow, Scikit-learn
Pandas, NumPy, Matplotlib,
Seaborn PROJECT EXPERIENCE
SQL, NoSQL and vector databses
(MySQL, PostgreSQL, MongoDB, PolyAI Mentor — AI Mentorship Platform
Weavite) AI / Backend Engineer
Docker • Designed and implemented a production backend using FastAPI, PostgreSQL, Redis, and
Weaviate.
Git, GitHub
• Built an end-to-end RAG pipeline covering document parsing, chunking, embeddings,
Django, FastAPI, Flask semantic retrieval, and LLM response generation.
Working with APIs (OpenAI, • Implemented asynchronous document processing with queues, retries, worker health
Google, Telegram) monitoring, and configurable model quotas.
• Integrated local and cloud LLM providers through Ollama and OpenAI-compatible APIs.
Basic Linux, CLI • Implemented authentication and authorization using JWT, Google OAuth 2.0, guest
sessions, and RBAC.
• Containerized the platform with Docker and managed database migrations with
LANGUAGES SQLAlchemy and Alembic.

Ukrainian (Fluent) Tarilka — AI Food Photography Platform
English (B2 FCE certificated) AI / ML Engineer
• Developed an AI image-generation pipeline for transforming casual food photos into
consistent commercial-style restaurant photography.
C E R T I F I CAT E S • Designed a custom inference workflow based on open-source generative models,
enabling control over multiple stages of the image-generation process.
Participation on Data • Built the model-processing pipeline for image transformation, generation configuration,
and automated output processing.
Science Bootcamp
• Developed supporting product infrastructure including user accounts, authentication and
Linux basics from authorization, asynchronous processing jobs, result management, and deployment.
Prometheus
Dzerkalo — AI Fashion Studio
B2 First (Cambridge
AI / Backend Engineer
Assessment English)
• Developed backend infrastructure for an AI-powered fashion platform that generates
virtual model photography from product images.
• Integrated an external AI image-generation service via API and built the application
workflow around asynchronous generation jobs.
• Implemented user accounts, authentication and authorization, generation history, credit
management, and result storage.
• Connected AI inference services with the application backend and production
infrastructure.

DANYLO KHOMYSHYN
ML ENGINEER

EasyTrip — Personalized Route Optimization
Reinforcement Learning
• Developed a reinforcement learning system for personalized tourist route optimization.
• Designed the environment, state/action space, reward function, and training pipeline
using Deep Q-Networks (DQN).
• Modeled user preferences from interaction history to dynamically adapt route
CONTACT recommendations.

[відкрити контакти](див. вище в блоці «контактна інформація»)
RESEARCH PROJECTS
[відкрити контакти](див. вище в блоці «контактна інформація»)
Scinalyze — Skin Lesion Analysis
github.com/DanyloShi Computer Vision / Deep Learning
telegram.org/khomyshyn13 • Developed an image-based skin lesion analysis system using a hierarchical deep learning
classification pipeline.
• Built a two-stage approach that first classifies uploaded skin images as benign or
SKILLS potentially malignant, then performs fine-grained classification into common lesion
categories.
Python (OOP, algorithms, data • Trained and evaluated multiple deep learning models on an open-source skin lesion
dataset, comparing their performance for both binary and multi-class classification.
structures)
• Integrated the trained models into a Flutter application, allowing users to upload a skin
RAG pipelines, LLM integration image, select the affected body area, and receive the predicted lesion category.
PyTorch, TensorFlow, Scikit-learn
Real Estate Price Prediction
Pandas, NumPy, Matplotlib,
Machine Learning
Seaborn
• Compared multiple ML approaches including Random Forest, XGBoost, CatBoost, and
SQL, NoSQL and vector databses
MLP for real estate price prediction.
(MySQL, PostgreSQL, MongoDB, • Performed feature engineering, cross-validation, and bias–variance analysis to evaluate
Weavite) model performance and generalization.
Docker
Git, GitHub
Django, FastAPI, Flask EDUCATION
Working with APIs (OpenAI,
Google, Telegram) Lviv Academic Gymnasium 2016-2023
Basic Linux, CLI Since I studied at a gymnasium with a math focus, I really liked math
and participated in math olympiads, which helped me pass the NMT
with the best score.
LANGUAGES
IT-Step Academy 2016-2021
Ukrainian (Fluent)
I studied various areas of IT such as web programming,
English (B2 FCE certificated) computer application development, games, robotics. I have
taught Python, C++, Java, HTML, CSS, JavaScript.

C E R T I F I CAT E S Lviv Polytechnic National University 2023-present
Institute of Computer Sciences and Information Technologies
Participation on Data
“Artificial Intelligence Systems ” program
Science Bootcamp
Linux basics from
Prometheus
B2 First (Cambridge
Assessment English)