Profile reviewedEmail verified

Ivan Antonijevic

AI Agent Architect & Full-Stack Engineer · Multi-Agent Systems, MCP & Production AI Workflows

I architect and ship production-grade AI systems, multi-agent workflows, and custom control planes for teams moving beyond simple LLM prototypes. My focus is on the critical "Day 2" engineering of autonomous systems: multi-agent orchestration (Google ADK / LangGraph), MCP (Model Context Protocol) tool integrations, strict RBAC, cost/token telemetry (OTel), and deterministic privacy guardrails. I built MATE (an open-source multi-agent tree engine and control plane with visual building, eval suites, and embeddable widgets) and actively architect domain-specific agent integrations (like StyleGuard for rule extraction). I work closely with founders and product teams to design robust agent architectures, automate complex operational workflows, and deploy maintainable AI infrastructure quickly and reliably.

Availability: Limited availability
Location: Serbia
Languages: English · Serbian
Timezone: Europe/Belgrade

BuildFix or finishAutomation

Request this Operator

Share a private project brief. No customer account is required; this Operator remains a preference until OpsSeek confirms an introduction.

What Ivan has shipped.

Proof of Work leads this profile: the context, the contribution and the real work behind each project.

Build· 2026

MATE

A production-grade multi-agent orchestration engine and control plane featuring dual ADK/LangGraph runtimes, RBAC, MCP tool integration, and live visual configuration.

The context: Most agent frameworks (CrewAI, AutoGen, basic LangChain scripts) stop at simple prototypes and lack "Day 2" production essentials. Teams struggle with multi-tenant RBAC, live configuration without redeploying code, observability (OTel), token budget controls, compliance audit logs (EU AI Act), and reliable tool integration via MCP.
What was shipped: Built MATE, an end-to-end multi-agent control plane and execution platform: Multi-Engine Runtime: Seamless execution on both Google ADK and LangGraph (switchable via environment variables). Control Plane & RBAC: Dual-server architecture (Auth/Proxy on port 8000, Execution on port 8001) with enterprise OAuth 2.0 / SSO and role-based permissions. Visual Builder & Work Room: React Flow drag-and-drop hierarchy builder, code canvas (Pyodide WASM / Flutter execution), and live embeddable chat widget. Protocols & Integrations: Model Context Protocol (MCP stdio/SSE), A2A agent-to-agent communication, and Slack bot integrations with Block Kit support. Observability & Guardrails: OpenTelemetry traces per step/LLM call, token budget throttling, eval test suites, and PII/hallucination guardrails.
Operator contribution: Sole creator and lead architect. Designed the entire system architecture, implemented the core engine in Python (FastAPI/SQLAlchemy), developed the frontend dashboard (React Flow, Monaco/Ace editors), built the automated DB migration system across PostgreSQL/MySQL/SQLite, and authored full unit/integration test suites (130+ tests).

DockerNode.jsPostgreSQLReactSupabaseTypeScriptClaude Coden8nAPI integrations

View project Public repository

Skills and tools

DockerNode.jsPostgreSQLReactSupabaseTypeScriptClaude CodeCursorn8nAPI integrations

Working fit

Capacity: 16 hours / week

Could Ivan Antonijevic be right for your project?

Tell OpsSeek what you need with this Operator carried into the private brief as your preference.

Request this Operator