AI engineering · intelligent systems · software products

Building practical AI products that feel sharp, useful, and real.

I’m Allen Mirzoyan, a Computer Science graduate and software developer building at the intersection of AI, developer tooling, and product engineering. Through MirzoyAI, I focus on systems that reason better, automate real work, and turn complex ideas into usable software.

Selected work

Projects that show how I think and build.

A mix of AI systems, technical experimentation, and product-oriented engineering. The focus is less on hype, more on building things that work and teach something valuable.

01 · LLM Systems

AtlasLLM

A retrieval-augmented generation project designed to explore document ingestion, vector storage, semantic retrieval, and local-model workflows in a practical architecture.

  • Document chunking and retrieval pipelines
  • Embeddings + vector database integration
  • Local LLM experimentation and system design
View repository ↗
02 · Applied Analytics

Chess Personality / Playstyle Analysis

An experimental project analyzing how people actually play chess — exploring style signatures, time pressure, decision patterns, and performance trends across different phases of the game.

  • Playstyle profiling and archetyping
  • Mistake pattern analysis and phase breakdowns
  • Data-heavy experimentation with practical insights
Browse projects ↗
03 · Engineering Portfolio

Software Experiments

Ongoing work across automation, intelligent interfaces, systems programming, and product prototypes — all tied together by a bias toward learning fast and building tangible outcomes.

  • AI-assisted product ideas and prototypes
  • Developer tools and workflow automation
  • Cross-domain experimentation with strong technical curiosity
See full GitHub ↗
What I build

Tools, systems, and workflows around real AI use-cases.

I’m especially interested in building AI that does more than generate text — systems that retrieve, reason, orchestrate, and help people work better.

RAG Systems

Retrieval pipelines that ground answers in documents, structured data, and domain-specific context.

Agentic Workflows

Multi-step AI systems that can plan, call tools, and complete more complex tasks reliably.

Developer Tools

Internal tools and product ideas that make engineering and experimentation faster and more effective.

Applied AI Products

Product-oriented prototypes that connect AI capabilities with clear user value and usable design.

About

Curiosity with engineering discipline.

Who I am

I’m Allen Mirzoyan, a recent Computer Science graduate from Maastricht University. I like building software that sits between technical depth and practical usefulness — especially where AI can improve how products behave, how systems reason, and how people work.

MirzoyAI is my space for turning that interest into real projects: developer tools, AI-powered systems, experiments, and prototypes that help me learn fast and build credibly.

Python LLMs RAG Embeddings AI Agents Software Engineering Automation Product Thinking

Current focus

Exploring AI system design Building stronger intuition for retrieval, orchestration, and production-minded LLM applications.
Turning projects into portfolio proof Creating work that demonstrates both technical understanding and product-building ability.
Sharpening practical engineering Improving not just what I build, but how I structure, ship, and communicate it.
Contact

Want to build something smart together?

If you’re interested in AI products, technical collaboration, or just want to connect about software, feel free to reach out. I’m always open to good ideas, interesting work, and strong technical conversations.