Володимир

Volodya Khomut [открыть контакты](см. выше в блоке «контактная информация») GitHub LinkedIn

Current role: Middle Python Developer / AI Automation Engineer

Experience: 3+ years

Python AI LLM AI Agents RAG

LangChain Fine-tuning / LoRA MCP

Prompt Engineering n8n Automation

Web Scraping FastAPI Django REST API

Selenium GraphQL WebSockets

PostgreSQL Docker Python

Work experience

Python Developer / Automation Lead Jul 2024 - Mar 2026 (1y 8m)
VAVI (Crypto Trading Analytics Platform)
Python AI AI Agents Fine-tuning n8n Web Scraping Django

Fast API REST API WebSockets PostgreSQL Docker Kubernetes

Automation

Designed and shipped 3 production AI agents automating arbitrage trading
signals, news sentiment analysis, and risk management in real time.

Arbitrage Scout — monitored spreads across 10+ exchanges via
WebSocket/REST feeds, generating trading signals with 83% accuracy and
cutting manual trader monitoring time by 80%.

News Sentiment Analyzer — scraped 20+ sources (Twitter, Reddit,
Telegram, RSS), used a DistilBERT model fine-tuned on proprietary market-
news data to score sentiment and adjust signal confidence, cutting news-
reading time by 90%.

Liquidity Guardian — evaluated order-book depth and slippage risk to
prioritize and filter trading signals, reducing failed/unprofitable trades by
45%.

Orchestrated agent scheduling and data pipelines via n8n; used LightGBM
and Prophet for signal/time-series modeling, with Grafana and the ELK stack
for real-time observability.

Built the underlying async backend (FastAPI, REST API) with Redis as both a
caching layer and task-queue backend, PostgreSQL/InfluxDB for analytics,
and a WebSocket-driven Telegram alert pipeline for real-time threshold
notifications.

Developed a secure internal operations portal with a safe SQL execution
layer (DDL blocking, table allow-listing) to protect production data.

Automated the delivery pipeline with Docker/Kubernetes and CI image
builds, ensuring consistent environments across the team.

Result: +15% additional profit in the first month, 70% of traders' routine
tasks automated, full audit log for compliance.

Software & Automation Engineer Dec 2022 - June 2024 (1y 6m)
All-in-one tool for course creators

Python AI LLM Prompt Engineering Kubernetes Selenium

Automation

Architected modular LLM integrations using OpenAI and Claude API
adapters, implementing dynamic prompt templates and multi-provider
comparison logic.

Built an automated LLM quality gate within CI/CD pipelines (GitHub Actions /
GitLab CI) that triggers pipeline failure based on hallucination detection and
(
confidence scoring low/medium/high risk flagging).

Enhanced system reliability by deploying Kubernetes k8s) manifests for (
workloads and services, accompanied by detailed Runbooks and
architecture documentation.

Developed a comprehensive E2E UI Automation framework using Selenium
and pytest (Page Object Model), covering smoke and regression suites for
critical user flows.

Improved observability and troubleshooting by implementing structured
logging and creating Jira-style bug report templates with automated artifact
(
collection screenshots/logs) on failures.

(
Authored technical documentation architecture.md, API examples, and
troubleshooting guides), significantly reducing onboarding time and
improving system maintainability.

Skills

Soft Skills
Technical documentation & architecture description

Code review, refactoring, working with legacy code

English — Intermediate written) (

E
T CHNICAL SKILLS
Languages & Frameworks:


Python (FastAPI, Django, asyncio, pytest, Selenium)

(
TypeScript / React basic frontend)

SQL — complex queries, window functions, CTEs

AI / LLM:

OpenAI API, Claude API, OpenAI-compatible adapters

AI Agents, Multi-Agent Orchestration

RAG pipelines (Qdrant, ChromaDB — retrieval, rerank, generation)

LangChain, MCP (Model Context Protocol)

Fine-tuning (QLoRA, Unsloth), vLLM, Embeddings, Checkpoints

Prompt Engineering, hallucination detection, confidence scoring

(
n8n workflow orchestration for AI agents)

Web Scraping / Parsing (BeautifulSoup, Scrapy, Playwright)

Telegram Bot API — automated reporting, alerts, and LLM output delivery
layer

Databases:

PostgreSQL, SQLAlchemy, asyncpg

(
MongoDB, Redis cache / queues)

MySQL, ChromaDB, Qdrant vector stores)
(
(
InfluxDB time-series)

DevOps & Tools:

Docker, Docker Compose, Kubernetes

GitHub Actions, GitLab CI/CD

Prometheus, OpenTelemetry, structured logging, Grafana, ELK Stack

(
Power BI data visualization)

Jira, Confluence, Git

ML / Analytics:

(
LightGBM, Prophet signal & time-series modeling)

Transformers (HuggingFace) — DistilBERT fine-tuning for sentiment
classification

Testing:

Selenium WebDriver, Page Object Model

(
pytest unit / integration / smoke / regression)

E2E UI automation, CI artifact collection