Nikolas K.

I'm an AI engineer and full-stack developer with 15 years of software experience and a BSc in Biomedical Engineering (Kharkiv National University of Radioelectronics). I build AI agent systems and the web products around them, and I'm working on a PhD in acoustic signal processing. 1. AI agent systems (core focus) - Tool-calling copilots and agents on Microsoft Agent Framework and any OpenAI-compatible LLM, with AG-UI streaming and CopilotKit chat. - Agents that act through your existing APIs and permissions: finding data, launching long-running jobs, tracking them and resuming the conversation when a job finishes. - Production concerns: server-side conversation storage, turn leasing, multi-provider LLM failover, handling of malformed tool calls, audit trail, MCP servers. 2. Full-stack engineering - Frontend: Angular (standalone components, signals) as the main framework, plus Vue 3 and Pinia. - Backend: C# / ASP.NET Core, NestJS (TypeScript), SignalR real-time, OAuth2/OIDC. - Telegram Mini Apps: Web App API, bots, authentication, payments. - Data: PostgreSQL, MS SQL Server, MongoDB, MySQL. Azure SQL Database, Azure Cosmos DB, Azure Database for PostgreSQL. Schema design, migrations (EF Core), T-SQL, query and index tuning. - Delivery and quality: Azure (App Service, Azure SQL, Key Vault, Blob Storage), Docker, Linux, CI/CD, xUnit, Jest, Playwright. 3. Research (PhD in progress): acoustic onset detection and sound source localization with AI - Detecting the onset of acoustic events in noisy field recordings. - Localizing sound sources from microphone arrays by combining signal processing (TDOA, beamforming) with neural networks. - Working across the whole ML pipeline: data collection, labeling, spectrogram features, training and evaluation. 4. Embedded (additional skill) - ESP32 firmware (ESP-IDF, FreeRTOS, I2C/SPI/UART, BLE/Wi-Fi, OTA), low-power design, PCB design with AI assistance, sensor and microphone boards.