Svyatoslav K.
Rating
Language proficiency level
CV
Up-to-date CV
I am a Data Science engineer with 2 years of commercial experience, specializing in natural language processing (NLP), large language models (LLMs), and AI-based chatbot systems. I have practical experience in fine-tuning open-source models, building RAG pipelines, and deploying secure, scalable ML applications on AWS. My stack includes LangChain, LangGraph, Hugging Face, and FastAPI, as well as in-depth knowledge of databases and MLOps tools. I work at the intersection of language and data, creating intelligent systems that are efficient, secure, and production-ready.
Technical Skills
Programming Languages | Python |
RDBMS /DB | MySQL, PostgreSQL, SQLite, MongoDB |
Operating Systems | Windows, Linux |
Tools and Technologies |
|
Other |
|
Skills and abilities
Services
Writing
Portfolio
-
RAG system for a Japanese financial institution
AI & Machine Learning- Task: Develop an intelligent system for analyzing financial reports with zero hallucination levels for working with complex Japanese documentation.
- Result: Implemented a "Gap Analysis" agent that detects low confidence in responses and redirects inquiries to experts. The system is fully scalable due to a microservices architecture on AWS.
- Technologies: FastAPI, AWS Fargate, PostgreSQL (pgvector), Japanese NLP tokenization, LangChain.
-
Conversational AI: Personalized chat-bot (Persona Cloning)
AI & Machine Learning- Task: Create an AI agent for Digital Media capable of imitating the unique communication style of the author (100% accuracy of persona) for automating chats with fans.
- Result: The system supports 1000+ simultaneous real-time dialogues. The use of RAG and model fine-tuning has ensured the generation of responses that are indistinguishable from human ones.
- Technologies: Fine-tuning (Mistral), LangChain, LangGraph, AWS (SageMaker, Lambda, ECR), FastAPI.
-
LegalTech: Intelligent scraping and structuring of data
AI & Machine Learning- Task: Build a system for automatic data extraction from unstructured lawyer profiles and a scalable pipeline for scraping.
- Result: Optimization of asynchronous parsers accelerated data collection by 200%. Implementation of OpenAI Batch API reduced text processing costs by 3 times while maintaining high accuracy.
- Technologies: Python, Scrapy, Celery, OpenAI Batch API, Docker, MongoDB.
-
SQL Agent: Private database analytics (Text-to-SQL)
AI & Machine Learning- Task: Create a secure system that allows users to retrieve data from PostgreSQL by asking questions in natural language (without knowledge of SQL).
- Result: A local solution based on open models has been implemented, ensuring complete data privacy. An intuitive interface on Streamlit allows non-technical specialists to independently generate complex reports.
- Technologies: LangChain Agents, LangGraph, Ollama (Open Source LLMs), PostgreSQL, Streamlit, Docker.
-
Voice AI agent for sales automation
AI & Machine Learning- Task: Automate lead qualification using voice AI with minimal latency.
- Result: Achieved real-time latency of less than 0.5 ms. Operational costs reduced by 30%, and lead conversion increased due to instant processing of requests.
- Technologies: Twilio WebSockets, ElevenLabs (Voice Synthesis), OpenAI, FastAPI, AWS.
-
45 USD Designs
Web DesignDesign for Conditioner
-
68 USD Prototype Site for Conditioner
Web DesignWebsite for the Conditioner. I performed in 3 days.
The buyer is pleased
Reviews and compliments on completed projects 4
27 April
180 USD
Stage 2: Scaling and Automation
Pleasant to collaborate
Very competent performer
![]()
30 January
361 USD
Stage 1: Core Prototype
Great performer
Qualitatively and responsibly performs tasks
Pleasant to collaborate
![]()
21 July 2025
293 USD
1. AI Agenda: Queue system
The project is completed according to the technical specifications
![]()
| Personal | Response review
5 July 2025
270 USD
Personal Project
The project was completed according to the technical specifications
![]()
| Personal | Response review
Activity
| Projects underway 1 | Budget | Added | Deadlines | Proposal | |
|---|---|---|---|---|---|
|
Stage 3: Integration (Telegram + Voice)
225 USD
|
| Latest proposals 10 | Budget | Added | Deadlines | Proposal | |
|---|---|---|---|---|---|
|
OCR system
|
|||||
|
Stage 3: Integration (Telegram + Voice)
225 USD
|
|||||
|
Stage 2: Scaling and Automation
180 USD
|
|||||
|
Stage 2: Scaling and Automation
180 USD
|
|||||
|
Stage 1: Core Prototype
361 USD
|
|||||
|
Development of an AI agent (RAG) for the corporate knowledge base
496 USD
|
|||||
|
TheJob
|
|||||
|
Personal Project
158 USD
|
|||||
|
Module .
180 USD
|
|||||
|
1. AI Agenda: Queue system
293 USD
|