Сергій

Serhii Kalyna
Backend Node.js Developer
[відкрити контакти](див. вище в блоці «контактна інформація»)

GitHub GitLab LinkedIn Telegram

SUMMARY
Backend Node.js developer with nearly 3 years of commercial experience building production systems end-to-end
— from REST APIs and real-time communication to job queues, AI integrations, and payment flows. Worked as
part of a team on the company's main product and as the sole backend engineer on client projects. Comfortable
owning a service from design to deployment, and quick to pick up new tools and technologies. Looking for new
opportunities in software development.

SKILLS
Backend: Node.js (Nest.js/Express.js), JWT, Redis, Socket.IO, Bull, SSE, PM2, Sentry, Docker Compose, OAuth2.0
+ payment + OpenAI integrations, AWS S3 + Cloudflare R2, E2E + stress tests
Database: MongoDB (Mongoose) , PostgreSQL (Sequelize ORM, TypeORM) + pgvector, MySQL
Integrations: Stripe, PayPal, TrustPay, WayForPay, NOWPayments
Web: HTML, CSS, JavaScript/TypeScript, Vue.js (Vuex), React (basic), Weweb

EXPERIENCE
Venstop.io
Backend Node.js Developer. April 2025 - June 2026 (14 months)
Worked on the company's main product, and additionally was the sole backend developer on two client projects.
"Sibrik.AI" — Telegram-bot platform aggregating 70+ AI models from multiple providers, allowing users to
generate videos, images, and audio (including music and speech) through a single interface. Includes credit-based
monetization, subscription plans with tiered access to models and a referral program.
The main tasks were:
AI provider orchestration (Wavespeed, KIE, FAL, and others) with polling, retries, and result normalization;
Asynchronous request processing via Bull job queues with per-provider concurrency;
Real-time result delivery via SSE backed by Redis Pub/Sub for cross-instance event propagation;
Payment integrations (Stripe subscriptions with webhook lifecycle, WayForPay, NOWPayments crypto)
The main stack is: Express.js, MongoDB (Replica Set), Redis, Bull, PM2.
"Dial Guard" — SaaS for automatic quality scoring of call-center operators via Binotel VoIP and external ML
service. Built webhook ingestion, job queue with retry/backoff logic and bootstrap recovery of unfinished calls,
multi-tenant architecture with per-tenant credentials, Stripe quantity-based subscriptions with proration, and
CRON-based CSV reports via Telegram. The main stack is: Express.js, MongoDB, Stripe, node-cron.
"Lviv blog" — city portal combining a content CMS with a cars and real-estate marketplace; split into two
backends to cover functionality outside Strapi's scope. Modeled 30 Strapi content types, built AI chatbot with
Google Gemini and pgvector semantic search with automatic vector indexing (RAG), real-time messenger and
notifications via Socket.IO, listing moderation, Stripe slot-based payments. The main stack is: Express.js,
MongoDB, Socket.IO, Strapi 5, PostgreSQL + pgvector.
 Meduzzen
Full-Stack JavaScript Developer. June 2024 - April 2025 (10 months)
Web application "AI Companion" for creating interactive AI-based virtual characters with personalized appearance,
personality, and communication style. The platform also supports generating images in different styles of the
created character in various scenarios using AI models. The application includes JWT authentication, payment
integration, and deployment pipelines for production usage.
The main tasks were:
The ability to create custom characters from present or custom characteristics of body and mood;
Message generation with chat settings and typewriter effect;
Photo generation including blurred photos, using character`s appearance and user`s wishes;
Payment system integration: Stripe, PayPal, TrustPay with recurring payments;
OAuth2.0 integration: Google, Discord, Twitter;
AI moderation of user messages to block unacceptable prompts;
Implementation of throttle on endpoints and container scaling to avoid overload + server stress testing;
Creating admin module for moderation custom characters and view purchase statistics;
The main stack is: TypeScript, Nest.js, PostgreSQL (TypeORM), AWS.

Training club
Backend Node.js Developer. October 2023 - June 2024 (8 months)
Web application "Training Club" for convenient sports activities, creating programs, exercises, learning the correct
technique with the help of professional trainers. Such features were developed: selection, saving and completing a
training program with specified exercises, approaches, repetitions, rest, media with correct execution of the
exercise and a stopwatch are also provided. The service also analyzes your trainings with compilation of statistics.
There is an opportunity to create own training programs and take students for training.
The main tasks were:
The ability to modify videos, including cropping, changing resolution, length and encoding into another format.
For this, media file processing libraries such as FFmpeg were used;
Pages have cursor pagination + lazy loading;
Queries to the database are often based on aggregations;
Saving, retrieving and deleting media from AWS S3;
Increased the coverage of E2E tests from 20% to 60%;
Code refactoring with the aim of increasing code readability, reduce code duplication and improving
testability;
Work with the file system for using server media files in tests;
The main stack is: TypeScript, Nest.js, MongoDB (Mongoose), React, AWS, Jest.

EDUCATION
Student of National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Information
systems and technologies, FICE.

LANGUAGES
English (B1-B2)
Ukrainian (Native)