Особисті дані приховані

Mykhailo Mykhailo Tymosheko
Tymoshenko Kyiv, 01030
[відкрити контакти](див. вище в блоці «контактна інформація»)
23/06/2008 — date of birth
AI/ML Engineer Mykhailo Tymoshenko (LinkedIn)
MykhailoTymoshenko08 (GitHub)

ㅡ
Summary Ambitious developer and Computer Science student with 2.5+
years of experience in web development and specialization in
AI/ML. I have practical experience in LLM studies, Computer
Vision system implementation and building complex RAG
architectures. I combine technical expertise with team
management and system design skills.

ㅡ
Technical Skills ● AI & Machine Learning: NumPy, Pandas, Matplotlib, Pydantic,
PyTorch, Scikit-learn, XGBoost, LangChain, RAG, OpenAI API,
YOLO, LSTM, Computer Vision (CV), ChromaDB, Tavily
● Programming Languages: Python, JavaScript(ES6+), C++,
Assembly, SQL
● Backend & Tools: FastAPI, Streamlit, Docker, CI/CD (GitHub
Actions), PostgreSQL, n8n, Git, Arduino
● Web Development: Next.js TypeScript, React.js, HTML, CSS,
JavaScript (ES6+), Figma, Canva, Angular21, Node.js,
Express.js
● CS Foundations: Analysis of Algorithms, Discrete
Mathematics, Object-Oriented Programming (OOP),
Computer Networking, Clean Architecture, Clean Code,
System Design
● Data Handling: CSV, JSON, psycopg2, pydantic, asyncio
● Soft Skills: Team Leadership, Problem Solving, Responsibility,
Communication, Rapid Learning, Intercultural Communication,
Project Management, Product Management

ㅡ
Experience RZK Soft-team / Founder & Lead Fullstack
June 2025 - PRESENT, Kyiv
Leading a multidisciplinary team of 8 specialists (AI, Mobile,
Frontend/Backend, Design, IoT) to deliver end-to-end software
solutions. We specialize in building smart AI-driven platforms,
cross-platform mobile apps, and complex hardware-software
ecosystems from scratch to deployment
Logistick AI Hackatone(Nova Poshta) / Finalist & Team Lead
March 2026, Kyiv
 Selected as a finalist to present a complex hardware-software
ecosystem on the main stage for Nova Poshta and Vodafone R&D.
Engineered an end-to-end system utilizing YOLO-based Computer
Vision for 1,100+ cameras and BLE for indoor asset tracking. Our
solution for real-time bottleneck detection and throughput
optimization is currently under discussion with the organizers for
potential strategic integration

ㅡ
Education Taras Shevchenko National University of Kyiv (FIT)
September 2025 - PRESENT, Kyiv
Bachelor of Computer Science (Specialty: F3)
Relevant Coursework: Analysis of Algorithms, Discrete
Mathematics, OOP, Assembly, Machine Learning.
Technical Lyceum of NTUU “KPI”
June 2022 - June 2025, Kyiv
Focus on advanced mathematics and physics.

ㅡ
Languages ● Ukrainian — Native
● Russian — Native
● English — Upper-intermediate (IELTS Academic B2 level)
● Spanish — B1
● Polish — A2

ㅡ
Projects
● Fraud Detection AI-Agent (OpenRouter Hackathon)
Designed a 3-layer data processing pipeline to identify
identity theft across scattered datasets (transactions, GPS
logs, SMS) while optimizing the LLM token budget.
Built a rule-based pre-filter using Python and Pandas to
eliminate 95% of normal daily transactions.
Implemented context enrichment utilizing the Haversine
formula for GPS data and text triggers for SMS to generate
structured dossiers.
Integrated LangChain, OpenRouter APIs, and Langfuse to
trace LLM evaluations, monitor token spend, and analyze
logical gaps in real-time.
● DreamHouse-AI
Analytical project in the field of machine learning for
predictive assessment of real estate value. The system is
based on regression models that analyze a set of key
parameters of objects and predict their market price.
● Customer Support AI Analyzer (Skelar Project)
Created a system to generate and evaluate complex
client-agent dialogues with specific personas (e.g.,
impatient, hidden_dissatisfied).
Developed specialized algorithms for sarcasm detection
and hidden dissatisfaction analysis, providing automated
service quality metrics.
● AdvancingBot - AI Mentorship Telegram Bot
Built a fully asynchronous Telegram bot for dynamic task
tracking using Python 3.11, Aiogram 3.x, and PostgreSQL
(SQLAlchemy + asyncpg).
Integrated OpenRouter API to analyze user reports and
generate adaptive feedback.
Implemented background tasks via APScheduler and
deployed the Docker-containerized application with
secure Pydantic-settings on an Oracle Cloud ARM64
server.
● AI Fusion Chat (Multi-LLM Judge): Engineered a
production-ready system utilizing an "LLM-as-a-Judge"
architecture to aggregate and validate responses from
multiple AI providers in parallel. Implemented RAG for
document processing and Tavily Search API for real-time
web grounding.
● Logistics AI Optimization (Nova Poshta Hackathon):
Developed an end-to-end monitoring system for a major
logistics terminal analyzing video feeds from 1,100+
cameras in real-time. Integrated Computer Vision (YOLO)
models to detect workflow bottlenecks and safety
violations, significantly increasing terminal throughput
● Customer Support AI Analyzer (Skelar Project):
Created a system to generate and evaluate complex
client-agent dialogues with specific personas (e.g.,
impatient, hidden_dissatisfied). Developed specialized
algorithms for sarcasm detection and hidden
dissatisfaction analysis, providing automated service
quality metrics.
● RZK Soft website: Designed and deployed a
high-performance website for RZK Soft. Focused on clean
code architecture and responsive design to ensure a
professional digital presence for the company