I build practical AI systems that go beyond model calls — combining LLMs, multi-agent orchestration, RAG, real-time applications, backend engineering, and production-style infrastructure.

My current focus is designing AI-powered products that are grounded, observable, maintainable, and built around real engineering workflows.

• 🤖 Focused on Artificial Intelligence, LLM Engineering & Software Engineering
• 🧠 Building with Generative AI, RAG, Multi-Agent Systems & Machine Learning
• 🔎 Interested in private knowledge systems, semantic retrieval, and intelligent assistants
• ⚙️ Exploring production AI engineering through APIs, real-time systems, persistence, observability, and containers
• 💡 Turning AI concepts into complete end-to-end software products
• 📚 Continuously improving my skills through hands-on engineering projects

A local-first, production-focused multi-agent AI platform built to explore the engineering required around modern LLM applications beyond a basic chatbot.

Orbyntiq combines a LangGraph supervisor workflow, specialist agents, private-document RAG, Model Context Protocol (MCP) capabilities, local LLM inference, persistent execution history, real-time WebSocket events, and a complete observability stack.

Highlights:

• Supervisor-driven routing across Research, General, and MCP specialist workflows
• Private knowledge retrieval using Qdrant vector search and grounded RAG
• Local inference and embeddings through Ollama
• Async FastAPI backend with an Angular 22 frontend
• Real-time LLM streaming and multi-agent workflow events over WebSockets
• Runtime state and caching with Redis, execution persistence with MongoDB
• Metrics and tracing with Prometheus, OpenTelemetry, Tempo, and Grafana
• Production-style Docker Compose, CI, security scanning, and load testing
• 349 automated backend tests passed in the current verified project snapshot
• Lightweight REST and WebSocket workloads validated with 100 concurrent local users and 0 failures

Technologies: Python · FastAPI · Angular · TypeScript · LangGraph · Ollama · RAG · Qdrant · Redis · MongoDB · MCP · WebSockets · Docker · Prometheus · Grafana · OpenTelemetry

➡️ Explore Orbyntiq

AI-powered intelligent scheduling assistant designed to understand natural-language requests and help users manage their schedules.

Technologies: React Native · Expo · Firebase · OpenAI · Speech-to-Text · TypeScript

➡️ Explore SophiAI

A collection of practical machine learning and neural network projects demonstrating model development, training, evaluation, and problem solving.

Repository coming soon.

Object-oriented Java applications demonstrating software architecture, OOP principles, GUI development, and problem solving.

Repository coming soon.

I'm interested in opportunities and collaborations involving:

Artificial Intelligence · LLM Engineering · Multi-Agent Systems · RAG · Machine Learning · AI-powered Applications

⭐ Building intelligent systems that connect AI models with real software engineering.