Kotlin · Android · Backend · Distributed Systems
I build Android applications, Ktor backends and distributed systems with clear architecture, consistent data flows and production-oriented observability.
I build software where straightforward CRUD logic is no longer enough: unreliable networks, concurrent workloads, secure code execution and consistent data states.
My main stack is Kotlin—from Android and Compose Desktop to Ktor backends powered by PostgreSQL, Redis and Docker.
I care about explicit architecture decisions, testable domain logic and systems that remain predictable when failures occur.
Selected projects focused on architecture, data consistency, scalability and real-world validation. Detailed case studies are available for the main projects.
Full Stack / AI-Powered Learning Platform
Full-stack platform for project-based software learning featuring a context-aware AI tutor, adaptive learning engine, integrated multi-file workspace, secure code sandbox, and centralized AI management supporting multiple AI providers.
Backend / Scalability
Kotlin, Ktor, PostgreSQL, Prometheus, Grafana, k6
A reservation backend designed to process concurrent booking requests in a controlled manner. It combines idempotent requests, transactional locking and asynchronous processing. k6 load tests together with Prometheus and Grafana make throughput, latency and bottlenecks observable.
Platform Engineering / Distributed Systems
Distributed multi-language sandbox execution platform
A platform for executing untrusted code in isolated Docker containers. Redis-backed queues, horizontally scalable workers, live SSE output, retries and dead-letter handling provide controlled processing and transparent failure states.
AI / Tooling
Semantic context storage and modular agent architecture
A coding agent for code and architecture questions with semantic long-term context. UI, orchestration, LLM integration, tool registry and memory service are separated so models and tools can be extended or replaced independently.
Offline-First / Sync
Synchronization architecture for unreliable networks and multiple devices
A synchronization system for unreliable networks and multiple devices. Local persistence, the outbox pattern, idempotent push and cursor-based pull keep state consistent across retries, connection failures and concurrent changes.
Android App
Android app for shift planning, deterministic time logic and analytics
An Android application for shift and work-time tracking with deterministic time calculations, StateFlow-based UI synchronization and persistent SQLite storage. It also includes monthly navigation and Canvas-based visual analytics.
Private case study · Details available on request
Desktop / AI Integration
Local AI integration with Clean Architecture and SQLDelight
A project management system with local LLM execution and a privacy-oriented architecture. Intent parsing translates prompts into concrete use cases and connects them directly to persistence and domain logic.
Private case study · Details available on request
Desktop App
Compose Desktop, SQLite, Exposed ORM
A desktop application for managing and billing operating costs. It focuses on type-safe data access, modular domain logic, live calculations and PDF generation for archive-ready statements.
Kotlin, Java, Go, Python, JavaScript, TypeScript, SQL, HTML, CSS
Ktor, Exposed ORM, Flyway, PostgreSQL, Docker CLI Integration, REST APIs, Clean Architecture
React, Vite, Tailwind CSS, Monaco Editor, Lucide Icons, Responsive UI
Docker, Docker Compose, Nginx, Container Isolation, Resource Limits, Linux Security Options
Sandbox Execution, CPU Limits, Memory Limits, PID Limits, no-new-privileges, Capability Dropping
Git, GitHub, IntelliJ IDEA, VS Code, Gradle, npm, Docker Desktop
Interested in working together or discussing Kotlin, backend and distributed systems? You can reach me here:
GitHub: github.com/gelovI
LinkedIn: linkedin.com/in/ivan-gelov
