Владислав

Deliar Vladislav
Python Developer
•​ Phone: [открыть контакты](см. выше в блоке «контактная информация»)
•​ Email: [открыть контакты](см. выше в блоке «контактная информация»)
•​ Telegram: @Vladde1231
•​ GitHub: github.com/VladDeliar
•​ LinkedIn: [открыть контакты](см. выше в блоке «контактная информация»)
•​ Location: Kolomyia (Remote work)

PROFESSIONAL SUMMARY
Technician Programmer with experience in developing, testing, and debugging software solutions based
on algorithms and mathematical models. Worked on data processing, program logic implementation, and
selection of appropriate programming languages. Designed input/output data structures, prepared test
cases, performed code reviews, and maintained existing software systems. Participated in software
automation, optimization of computational processes, and technical documentation preparation.

TECHNICAL SKILLS
Python, FastAPI, Django, PostgreSQL, MongoDB, Redis, Google Gemini API, OpenAI API; prompt
engineering, structured (JSON) output, validation, n8n, Celery, APScheduler, aiogram (Telegram Bot
API), Google Sheets / Drive APIs, web scraping (BeautifulSoup, Selenium), Docker Compose, Git

COMMERCIAL WORK EXPERIENCE
Junior Software Engineer | Binotel April 2025 – April 2026
Stack: Python, FastAPI, PostgreSQL, SQLAlchemy, Pydantic , Redis, Docker, Git
•​ Developed a production-ready Point-of-Sale (PoS) backend system from scratch using Python and
FastAPI, with a focus on scalability, maintainability, and security. Leveraged async/await (asyncio) for
non-blocking I/O and Pydantic for strict data validation and serialization. Designed a normalized
relational schema in PostgreSQL, managed via SQLAlchemy async ORM and versioned with Alembic
migrations. Implemented JWT/OAuth2 authentication with role-based access control (RBAC). Used
Redis for caching and rate limiting, Celery for background task processing (receipts, reporting), and
containerized the full stack with Docker and Docker Compose. Maintained high code quality through
pytest test suites, and automated CI/CD via GitHub Actions. APIs documented with OpenAPI 3.0 /
Swagger UI.

PET PROJECT PORTFOLIO
1.​ PoS (Point of Sale)
Stack: Python, FastAPI, MongoDB, Redis, Pydantic, JWT / OAuth2 / 2FA, Jinja2 + Alpine.js
•​ Built a cloud POS/ERP SaaS platform with multi-branch support: catalog & menu, orders, KDS,
inventory, delivery, loyalty, finance, and analytics.
•​ Implemented offline-first sync with idempotency (no duplicate orders), real-time updates via
WebSocket + Redis, hierarchical RBAC with JWT/2FA, and a public OAuth2 API with webhooks.
2.​ olx-domria-scraper
Stack: Python, Django, BeautifulSoup, Playwright, Pydantic, aiogram, Celery + Redis, PostgreSQL / SQLite,
pytest,Docker, GitHub
•​ Scrapes listings from OLX and Dom.ria, stores them in a DB, serves a web UI + REST API, runs
scheduled scrapes (Celery), and pushes Telegram alerts matching user subscriptions (district / price /
rooms).
•​ Parses embedded JSON state instead of fragile CSS selectors, validates with Pydantic, upserts
idempotently, detects price anomalies, and dedups notifications; 32 pytest cases + CI.
3.​ ai-call-audit
Stack: Python, Google Gemini API (transcription + analysis), Google Drive API (OAuth2), openpyxl
•​ A bot that pulls call recordings from Google Drive, transcribes them via Gemini, and scores manager
performance against set criteria (1/0 scores + comment), uploading the finished Excel report back to
Drive.
•​ Failed calls are auto-highlighted in red; re-runs skip already-processed files.

4.​ ReorderRadar (B2B)
Stack: Python, aiogram, FastAPI, OpenAI, Google Sheets API, SQLAlchemy, APScheduler, Docker, pytest
•​ Built a multi-tenant Telegram bot with a backend for B2B distribution: it detects customers who are
overdue for a repeat order, sends the manager a daily call list with a personalized (OpenAI) call script,
records contact outcomes, and emails the owner a short evening analytics summary.

5.​ Ad-Intelligence-Workflow
Stack: Python, Meta Ad Library API, Gemini, Pydantic, Google Sheets, APScheduler, Docker
•​ Built a repeatable AI pipeline for competitive intelligence: collects competitors' active ads (D2C
skincare, UK + Germany markets), uses an LLM to break each creative into a structure (hook / offer /
CTA / angle / format / language), and produces a clean Google Sheets report.
•​ Implemented 3-level deduplication (by ad_id, text, and URL) and AI quality control: a strict Pydantic
schema with strict JSON, a confidence < 0.7 → needs_review routing rule, a "don't hallucinate"
prompt, and a ~10% manual spot-check logged to validation_log.csv.

6.​ gravity-automation-task
Stack: n8n (self-hosted, SQLite), Google Gemini API, Google Sheets API, Telegram Bot API, WordPress +
Elementor, Docker Compose
•​ Built end-to-end automation that takes a website form submission through the full pipeline with no
manager involvement: AI classifies the lead, data is logged to a spreadsheet, and the manager is
notified in Telegram.
•​ Implemented the n8n workflow: receives the form webhook, classifies the lead via Gemini
(professional / wireless / consultation), logs every lead to Google Sheets, sends a Telegram
notification with the suggested next step, and returns a response to the form.
•​ Deployed the full stack with Docker Compose (n8n + WordPress)

LANGUAGES
•​ Ukrainian — Native (fluent)
•​ English — B2