AI Systems Builder | AI Automation Engineer | LLM & Voice Systems
I build working AI systems that combine local LLMs, persistent memory, voice interaction, and automation workflows.
I use AI-assisted development workflows to rapidly prototype, integrate, test, and ship practical AI systems for support, automation, and conversational use cases.
Featured Demo: Nyx
Nyx - Local AI Memory & Voice System
Nyx is a locally running AI system with persistent memory, real-time interaction, voice output, image workflows, and model switching. It demonstrates practical LLM integration, memory-based interaction, and AI automation workflows.
AI Systems & Delivery Work
Local AI Setup
Independent local AI testing environment. Runs and evaluates model groups for reasoning, coding, lightweight tests, and Nyx-specific memory behavior without relying only on hosted APIs.
Voice Assistant Call Qualifier Bot
AI voice automation for real customer call workflows.
- Built voice automation for HVAC/AC call qualification using Twilio, Flask, Whisper STT, TTS pipelines, SQLite, and local LLM fallback.
- Designed Serbian/Croatian conversational flows, intent handling, fallback logic, and interrupt handling for support use cases.
- Iterated with client/business feedback toward real-world call handling.
Voice Prime CRM
AI-assisted CRM and call-center workflow system.
- Built AI-assisted workflow system integrating Asterisk PBX, Flask API, AGI server, lead tracking, call history, analytics, GUI, and dashboard.
- Connected voice operations, CRM data, and workflow automation into a usable support/sales prototype.
Klima Servis Website
Client website / practical delivery project. Delivered a responsive commercial website for a local HVAC business with direct client communication, form handling, SEO, PWA behavior, and performance-focused frontend implementation.
How I Build AI Systems
LLM
Systems
- Local models / Ollama
- Qwen / DeepSeek
- RAG patterns
- FAISS vector search
- Memory systems
AI-Assisted
Development
- Prompt-to-code workflows
- Iterative debugging
- Rapid prototyping
- System integration
- Testing and refinement
Voice
AI
- Whisper STT
- Kokoro / Edge TTS
- Twilio
- Asterisk PBX
- Conversational flows
Automation
Workflows
- CRM systems
- Support automation
- Call qualification
- Customer-facing AI tools
- Workflow dashboards
Systems
Layer
- Python
- Flask APIs
- SQLite
- JavaScript / HTML / CSS
- Local hardware optimization
Remote Work & Setup
Remote-ready AI systems development with support, operations, and client delivery experience.
Remote-First Delivery
5+ years of remote work experience with international teams and clients. Comfortable with async updates, documentation, requirements clarification, and independent execution.
Support & Operations Context
Background across support and technical operations helps me build AI tools for real workflows, not just demos: call handling, customer support, CRM data, and handoff logic.
Equipped Home Office
Fully equipped home office for remote AI automation, LLM integration, testing local models, voice workflows, and collaborative delivery.
Prototype to Working System
I move from problem definition to integrated prototype quickly, then iterate with feedback until the workflow is understandable, testable, and usable.
Research Notes
AI Memory, Identity, and Local Model Behavior
I also explore long-term questions around AI memory, identity, emotional modeling, and local model behavior. These experiments support my practical work on conversational AI systems, but the main focus of this portfolio is working AI automation.
Contact
Currently Open To
- AI Automation Engineer
- AI Systems Builder
- LLM Integration Specialist
- Conversational AI Developer
- AI Support / Solutions Engineer
- Remote AI workflow roles