Юрій

Yuri Kasian
[открыть контакты](см. выше в блоке «контактная информация») | [открыть контакты](см. выше в блоке «контактная информация») | [открыть контакты](см. выше в блоке «контактная информация») | Ukraine
(Remote)
Professional Summary
AI Solution Architect and Full-Stack Engineer with 7+ years of industrial experience,
specializing in architecting and deploying engineering-grade AI systems and
robust Node.js-based automation. Expert in multi-agent LLM orchestration, RAG
validation pipelines, and event-driven module architectures for transforming messy,
manual processes into structured, repeatable systems. Proven track record of accelerating
content production by 16x and achieving 0.98 output accuracy, directly translating to faster
lead response, smarter estimates, and reliable workflows. Adept at leveraging
modern JavaScript (ES6+, Node.js), n8n, and diverse LLM models like Claude 3.5
Sonnet to drive significant business impact and technical innovation in practical, real-world
applications.
Hard Skills Matrix
AI & LLM Orchestration:
• Architectural Patterns: Multi-Agent Pipelines, Retrieval-Augmented Generation
(RAG), Decision Engines (Structured Rules-Based AI), Event-Driven Module
Architecture, API Gateway & Data Normalization
• Models: OpenAI (GPT-4o), Anthropic (Claude 3.5 Sonnet), DeepSeek (R1, Reasoner,
Chat), Groq / Llama-3, Gemini, Copilot
• Prompt Engineering Systems: Master Prompt System (v5/v6), Authority Engine
(v2.11 + PLATINUM), RFID OS Prompt Operating System (v4.0), Checklist
Framework (3.0)
• Techniques: Chain-of-Thought, Few-Shot Reasoning, System Instruction
Architecture, Multi-Agent Role Orchestration, Version-Controlled Prompt Chaining
• Anti-Hallucination: Anti-Fantasy Principle, RAG Validation Pipeline, Multi-Agent
Checks
Automation & Workflow Systems:
• Core Automation Engine: n8n (Advanced Custom JS Nodes)
• Workflow Platforms: Zapier, Telegram Bots
• API Patterns: JSON-LD Schema, LocalStorage Persistence, Cross-Frame
Communication, LLRP Integration
Programming & Core Technologies:
• Expert: JavaScript (ES6+, Node.js), HTML5/CSS3 (Semantic Layout)
• Advanced: SQL (Database Queries)
• Intermediate: Python (Data Processing, Scripts)
Industrial & Systems Thinking:
• Domain Expertise: UHF RFID (Cold Chain, Warehouse Asset Management, Aviation
Equipment Tracking, Tire Marking Systems, Garment Tracking)
• Hardware Specifications: Temperature (-40°C to +200°C), IP Rating (IP68),
Materials (Ceramic, Metal, PET)
• Standards: ISO, EN, ETSI, GS1, DTU, SBI, DSTU, ASHRAE, VAT, VDA
Quality Assurance & Performance:
• Validation: Zero-GPT Detection (<5% AI Score), Originality.ai (<10% Similarity),
Writer.com Detection (96%+ Human), Fact-Checking Framework
• Metrics: Content Production Speed (16x faster), Output Accuracy (0.98), SKU
Processed (10,000+)
Frontend & UI/UX:
• Capabilities: Clean Code CMS Approach, Print-Optimized Rendering, Dark Theme
UI, Form System Generation (Checkboxes, Inputs, Validators), Modal Dialog
Framework
• Components: RFID Config Form Generator, Inline Equipment Calculator
Detailed Experience
AI Solution Architect | Marketplace AI Hub & DocFlow Integration
• Architected and deployed a multi-agent AI system for automated technical
documentation generation, leveraging Custom JS within n8n as the core automation
engine, transforming messy, manual processes into structured, repeatable systems.
• Engineered RAG Validation Pipelines with multi-agent checks to minimize LLM
hallucinations, achieving an output accuracy of 0.98 and enforcing the Anti-Fantasy
Principle for data integrity and consistent outputs.
• Orchestrated diverse LLM models including GPT-4o for structured
generation, Claude 3.5 Sonnet for technical synthesis and validation, and DeepSeek
(R1/Reasoning) for analytics and complex logic, demonstrating strong tool judgment.
• Achieved a 16x reduction in content production time (from 4 hours to 15-20
minutes) for technical articles across 15+ languages, directly enabling faster lead
response and proposal drafting.
• Developed a sophisticated Multi-Agent DocFlow pipeline for automatic recognition,
classification, and structuring of incoming technical documentation, building outputs
like job summaries and ensuring compliance with Google E-E-A-T.
• Managed and processed over 10,000+ SKUs and structured content across 18
industry clusters, showcasing robust systems thinking and scalability for complex
business operations.
AI Orchestrator "Asymmetric Logic" v3.2 + Global Model Registry
 • Architected a high-precision meta-orchestrator capable of autonomous decision-
making across a curated database of 192 AI models, demonstrating advanced systems
thinking for AI automation.
• Developed a dynamic routing engine that treats a 192-model database as a
"contractor marketplace," selecting the optimal model based on an asymmetric scoring
model (0.98 accuracy threshold) for each sub-task, ensuring consistent and reliable
workflows.
• Implemented economic scalability by utilizing the full 192-model spectrum,
achieving up to 85% cost reduction by offloading high-volume tasks to "Tier 3"
models while preserving "Tier 1" reasoning for critical architectural layers, directly
contributing to smarter estimates and better job costing.
• Curated and structured a proprietary dataset of 192 global AI models (OpenAI,
Anthropic, DeepSeek, Google, Meta, open-source providers), cataloged by 12+
parameters including Pricing, Latency, Tier, and Domain Strengths, enabling
continuous improvement and iteration.
Technical Content Systems Architect | rfid.org.ua
• Implemented a comprehensive Prompt Operating System (RFID OS
v4.0) and Master Prompt System (v5/v6) for complex engineering tasks and content
authority verification using the Authority Engine (v2.11 + PLATINUM), showcasing
daily, hands-on AI use.
• Developed a robust Quality Assurance Framework incorporating Zero-GPT
Detection (<5% AI Score), Originality.ai (<10% Similarity), and Writer.com
Detection (96%+ Human) to ensure content authenticity and quality, improving
consistency of AI-generated outputs.
• Created dynamic UI components such as an RFID Config Form Generator and
an Inline Equipment Calculator using JavaScript (ES6+) for real-time technical
specifications and cost calculations, directly supporting sales and operations teams.
Education
Master of Engineering | Artificial Intelligence and Automation Systems