DEV Community

Aamer Mihaysi
Aamer Mihaysi

Posted on

Hello, I'm Aamer — AI Engineer Building Agentic Systems

#ai #llm #agentic #rag

Hello, I'm Aamer — AI Engineer Building Agentic Systems

👋 Welcome to my corner of the dev community!

I'm Aamer Mihaysy, an AI Engineer and Solutions Architect focused on making Large Language Models actually useful in production. If you're into agentic systems, RAG pipelines, or just figuring out how to deploy AI without breaking everything, we're probably going to get along.

What I Work On

My day-to-day revolves around three core challenges:

🤖 Agentic Orchestration

I'm building multi-agent swarms that coordinate to solve complex tasks. Think: ephemeral agents that spin up, do their job, and vanish. Local-first, privacy-focused, using open-source LLMs (Ollama ftw).

Key tools: LangGraph, LangChain, custom orchestration layers

🧠 RAG That Actually Works

RAG is easy to demo. Hard to productionize.

I'm building graph-augmented retrieval systems that go beyond flat vector search — extracting concept networks, traversing relationships, and giving LLMs the structure they need for complex reasoning.

Current focus: Knowledge graphs from unstructured text, hybrid semantic + structural retrieval.

🛡️ AI-Powered Security

Building an AI SOC that uses LLMs for threat hunting, alert triage, and incident response. Real multi-agent concurrent processing with knowledge capsules for institutional memory.

📊 Adaptive Learning Systems

Also working on personalization engines that track mastery, identify struggle patterns, and adapt content paths in real-time.

My Philosophy

Production over demos. Anyone can get GPT-4 to say something impressive. Getting Mistral 7B (local) to reliably extract structured JSON from messy text? That's engineering.

Open-source first. If I can run it locally, I will. Privacy matters. Dependencies should be opt-in, not mandatory.

Code is craft. Clean architecture > clever hacks. Systems that scale > systems that impress in a tweet.

Tech Stack

Languages: Python, TypeScript

Frameworks: FastAPI, Next.js, LangChain, LangGraph

Data: PostgreSQL, Prisma, DuckDB, Qdrant

ML/Ops: Ollama, MLflow, DVC, Docker

Visual: Streamlit, Plotly

What's Coming

On Dev.to, I'll be writing about:

  • Building production RAG systems — the hard parts nobody talks about
  • Agentic architecture patterns — what works, what doesn't
  • Local LLM deployment — getting open-source models to production quality
  • AI system design — from prototype to production-ready

Expect code, architecture diagrams, and honest post-mortems.

Connect With Me

What I read from you:

If you're working on:

  • Multi-agent systems
  • Graph RAG / knowledge extraction
  • Production LLM deployment
  • AI security / SOC automation

...let's chat. Always open to collaborating on interesting problems.

This is the start of a longer conversation. Looking forward to learning from this community.

#ai #llm #agentic #rag #mlops

Top comments (0)