Микола

MYKOLA BALIELOV
Senior Backend Developer | AI & System Architecture
Kropyvnytskyi, Ukraine | [відкрити контакти](див. вище в блоці «контактна інформація») | [відкрити контакти](див. вище в блоці «контактна інформація»)

PROFESSIONAL SUMMARY

Senior Backend Developer with over 5 years of experience architecting high-load Node.js/NestJS
systems, specializing in real-time AI integrations, cloud infrastructure, and distributed microservices.
Proven track record of designing provider-agnostic cloud environments, custom telephony protocols
(SIP/RTP), and scalable multi-tenant SaaS platforms that deliver measurable business and educational
value.

TECHNICAL SKILLS

Category Technologies & Tools

Languages & Node.js, TypeScript, NestJS (v10/v11), Fastify, Express.js, C++ (N-API)
Frameworks

AI & Real-time Vertex AI (Gemini Live, Imagen 3, Gemini 1.5/2.5 Pro/Flash), MCP SDK, SIP
Systems (UDP/TCP), RTP, WebSockets, Socket.IO

Data & Messaging PostgreSQL, Prisma ORM, Redis, MongoDB, BullMQ, Google Cloud Tasks,
ZeroMQ

Cloud & DevOps GCP (Cloud Run, Cloud Storage), AWS (EC2, S3, ECS, CloudWatch),
Docker, GitHub Actions, GitLab CI/CD, Jenkins

PROFESSIONAL EXPERIENCE

Lead Backend Developer | Under NDA
October 2025 – Present

● Real-time AI Voice Platform: Architected an end-to-end AI voice call platform integrating Google
Gemini Live with SIP/RTP telephony, enabling fully autonomous outbound sales calls with sub-200 ms
audio round-trip latency.

● Custom Telephony Stack: Implemented a custom SIP stack from scratch using Node.js UDP sockets
and built a C++ N-API native addon for RTP packet handling and PCM resampling, offloading
real-time processing from the JS event loop.
● Multi-tenant SaaS Architecture: Built a multi-tenant NestJS API using Prisma 6 client extension
chaining for transparent data isolation and audit logging, secured by a layered RBAC system.
● Cost Optimization & AI Configuration: Integrated Answering Machine Detection (AMD) to eliminate
wasted LLM tokens and designed a schema-driven (Zod) live-tunable AI config system for safe prompt
adjustments without redeployment.

Senior Backend Engineer | Nexis EdTech Platform
Early 2026 – Present

● Multimodal AI Content Generation: Architected a two-stage pipeline on Vertex AI (Imagen 3 +
Gemini 1.5 Pro) utilizing native gs:// references, enabling automated generation of playable sandbox
browser games from natural language.
● Cloud-Native Async Processing: Implemented a provider-agnostic queue and storage abstraction
layer (NestJS DI) that seamlessly switches between local environments (Bull/Redis/MinIO) and
production (Google Cloud Tasks/GCS).
● Performance & Output Engineering: Guaranteed type-safe AI outputs via Gemini's native schema
enforcement and built a deferred LLM-grading system with concurrent scaling constraints, reducing
request latency by ~60%.

Backend Developer | FinTech & B2B Solutions
May 2021 – Late 2025

● Real-time Systems & Messaging: Designed and developed a full-featured WebSocket-based chat
system with real-time messaging, tracking, and push notifications; implemented microservice queues
using ZeroMQ and Bull.
● Financial Integrations: Handled high-volume historical and real-time market data by integrating
premium financial APIs (Polygon, FMP) for live Forex and cryptocurrency pair tracking.
● API Orchestration & Payments: Engineered complex integrations for Zoom API automated
calendars and multiple globally recognized payment gateways (Stripe, PayPal, Apple Pay, Google
Pay).

EDUCATION

Master's Degree in Informatics / Computer Science
Volodymyr Vynnychenko Central Ukrainian State Pedagogical University, Kropyvnytskyi, Ukraine