Switch to English?
Yes
Переключитись на українську?
Так
Переключиться на русскую?
Да
Przełączyć się na polską?
Tak
  1. Freelancers

Kyrylo K.

Offer Kyrylo work on your next project.

Ukraine Kyiv, Ukraine
1 month 3 days back
Available for hire available for hire
on the service 1 month 12 days
Freelancer Kyrylo K. — Ukraine, Kyiv. Core skill — AI & Machine Learning, AI Consulting

Rating

Successful projects
No data
Average rating
No data
Rating
295
AI & Machine Learning
202 place out of 2836
AI Consulting
46 place out of 255

Skills and abilities

Programming

Administration

Writing

Portfolio


  • AI Support Agent - a company selling bus tickets

    AI & Machine Learning
    RAG system for customer support based on a knowledge base with company rules (returns, transfers, regulations). The central task is to relieve 10 live operators without losing service quality. RAG according to company rules: Vectorization of internal documentation with retrieval settings for short chat queries from customers. Complex prompt engineering for escalation: The agent independently decided when to transfer the dialogue to an operator. Developed a multi-level evaluation logic: type of problem, level of uncertainty, emotional tone, dialogue history -- minimizing false escalations is critical with a pool of only 10 operators. Python state machine: The finite automaton managed the stages of the dialogue (problem identification -> database search -> clarification -> resolution -> escalation decision -> closure). Each state defined the available set of tools and context format for LLM. Integration with internal API through Tool Calling: The agent in real-time received status and booking details, flight route (from/to, stops), current bus location, current schedule, and delays.

  • AI agent for legal consultations (Embeddings, academic project)

    AI & Machine Learning
    AI agent for automating legal consultations with vectorization of the legal corpus. Pipeline: parsing PDF court decisions and procedural documents, structured chunking, encoding with an embedding model considering the specifics of legal language. Semantic search through the vector index allows finding relevant precedents. Responses contain quotes from specific documents - critical for the legal context.

  • Voice AI agent for a real estate agency (RAG)

    AI & Machine Learning
    Built a RAG pipeline for an agent that autonomously called clients and consulted them. The main difficulty was scoring and ranking retrieved chunks in the context of a live call: implemented query rewriting, hybrid search (keyword + vector), re-ranking by relevance. Configured the embedding model on a corpus of documents about real estate objects. The agent escalated edge cases through HITL.

  • Naymi-ai

    AI & Machine Learning
    An open-source CLI agent based on AI, designed to automate and optimize the full cycle of job searching for IT professionals. Naymi AI levels the playing field against corporate AI filters, acting as a personal automated AI recruiter directly in the terminal.

    Key features and technical achievements:

    … Automated search and integrations: A portal scanner has been designed that integrates with the APIs of major applicant tracking systems (ATS) such as Greenhouse, Ashby, Lever, as well as job boards for automatic collection and filtering of relevant job postings.

    Intelligent matching system (Smart Match): An evaluation algorithm has been developed that analyzes job descriptions based on the candidate's profile, assigning a match score (A–F) based on skills, market salary, and identifying potential hidden risks ("red flags").

    Dynamic resume generation tailored for ATS: A headless Chromium-based engine has been created for instant generation of perfectly tailored, 100% ATS-compatible resumes in PDF format and cover letters for each specific role.

    Tracking and process automation: An embedded tool for managing the feedback funnel has been implemented, which tracks the schedule of follow-up communications and analyzes reasons for rejections.

    Interview preparation module: A feature has been created that automatically extracts real user experience and structures it into stories using the STAR method (Situation, Task, Action, Result), tailored to specific job requirements.

    Technology stack: AI agents (Claude Code / OpenCode), LLM, REST API, CLI, Headless Chromium, Markdown-based databases.

    #AI-agent

Activity

  Latest proposals 3
AI agent for optimizing business processes (pricing analytics)
248 USD
AI Agent
609 USD
Automation agent through Claude MCP
451 USD