Михаил

Mykhailo Kohutenko
[open contact info](look above in the "contact info" section) [open contact info](look above in the "contact info" section)
[open contact info](look above in the "contact info" section)

SUMMARY
AI & Infrastructure Engineer with a strong mathematical foundation (BSc in Mathematics & Computer
Science) and an ongoing MSc in Artificial Intelligence Systems. Experienced in designing production-grade
GraphRAG architectures, asynchronous Multi-Agent LLM systems, and deploying physical AI compute clusters
for enterprise environments. Proven track record of reducing LLM token consumption by 93x through
topological AST-based code analysis and automating L1/L2 engineering workflows via custom GenAI
microservices.

SKILLS
AI & LLM Engineering: Programming & Databases:
GraphRAG, LLM Orchestration, Gemini API, Groq Python (Core, OOP, asyncio, aiogram 3.x, AST), SQL
API, OpenRouter, Vector Embeddings, Prompt (SQLite, aiosqlite, Relational Schema Design), Graph
Engineering, Transformers, Hugging Face, Deep Databases (Neo4j, Cypher), JSON (REST APIs,
Learning (CNN, RNN, LSTM, GRU), Scikit-learn, Schema Serialization), NumPy, Pandas.
TensorFlow.
Mathematical Foundations:
Infrastructure & MLOps: Graph Theory, Linear Algebra, Mathematical
AI Cluster Initialization & Configuration, Server Statistics, Probability Theory, Optimization
Hardware Selection & Validation, Docker, Linux, Methods, Genetic Algorithms.
Git, Virtualization & Networking Fundamentals
(VMware, VLANs, TCP/IP).

PROFESSIONAL EXPERIENCE
AI & Infrastructure Engineer 10/2024 – Present
Alesta UA Kyiv, Ukraine
Enterprise GraphRAG Code & Configuration Analyzer (Internal R&D):
Architected an end-to-end GraphRAG microservice to analyze internal automation
•

scripts and enterprise repositories by parsing source code into Abstract Syntax
Trees (AST) and mapping structural dependencies into a Neo4j graph database.
Achieved a 93x reduction in LLM context token usage by combining topological
•

subgraph traversal (Cypher) with Gemini API vector embeddings and JSON
serialization, eliminating model hallucinations during code review and
diagnostics.
Enterprise Multi-Agent AI Copilot for Engineering Support:
Designed and deployed an asynchronous GenAI support assistant using Python
•

(asyncio, aiogram 3.x) to provide field and support engineers with instant access to
hardware troubleshooting protocols and technical documentation.
Engineered a fault-tolerant LLM routing layer integrating Groq API and
•

OpenRouter via JSON REST APIs, backed by an asynchronous SQL database
(aiosqlite) for session state persistence, audit logging, and dynamic role-based
prompt injection.
Integrated a real-time multimodal voice pipeline (edge-tts) and automated context
•

compression, maintaining sub-second response latency under concurrent
engineering team usage.
AI Cluster & Server Infrastructure Deployment:
Conducted hardware selection, initialization, configuration, and validation for
•

enterprise AI compute clusters and high-availability server solutions tailored to
B2B client requirements.
Assisted in physical server rollouts, system inspection, and basic
•

network/virtualization setup (OS installation, VLAN segmentation, basic
hypervisor configuration) alongside senior implementation engineers.
Structured technical documentation, architectural specifications, and analytical
•

reports to streamline communication with client CTOs and engineering
departments.

EDUCATION
MSc in Systems and Methods of Artificial Intelligence 09/2025 – 12/2026
Igor Sikorsky Kyiv Polytechnic Institute (KPI), IASA Kyiv, Ukraine
Master’s Thesis: Topological Code Analysis and Context Optimization using
•

GraphRAG, AST Parsing, and Vector Embeddings.
Applied Research: Implemented neural network optimization algorithms
•

(Backpropagation, Genetic Algorithms) and graph-based data structures in
Python.
BSc in Mathematics and Computer Science 09/2021 – 05/2025
V. N. Karazin Kharkiv National University Kharkiv, Ukraine
Core Focus: Linear Algebra, Mathematical Statistics, Optimization Methods, and
Probability Theory.

CERTIFICATES
Natural Language Processing Specialization Deep Learning Specialization
DeepLearning.AI & Stanford University (Built DeepLearning.AI & Stanford University (Engineered
sentiment analysis, NER, and transformer-based deep L-layer neural networks from scratch in
sequence models using RNNs, LSTMs, GRUs, and NumPy and CNNs in TensorFlow for image
Hugging Face). recognition).

Machine Learning Specialization
DeepLearning.AI & Stanford University
(Implemented supervised regression/classification
models with L1/L2 regularization and collaborative
filtering recommender systems).