• Building
  • Writing
  • About

AI engineer · founder · Curious Fella

Dmitriy Leybel

I work on knowledge systems and integrate language models into the cool things I build.

Right now that looks like DuxMedic, Stoica, Hippograph, and Pyllments. Consultancy through Prudent Patterns.

Work with me Read the writing

Now

Building DuxMedic. Advising through Prudent Patterns.
Refining Pyllments.

Elsewhere

GitHub · X · LinkedIn

About

Born in Kiev, raised in Los Angeles, now in Oregon.

The longer version.


Building

Things I am working on

My companies, projects, passion.

DuxMedic

Building

A medical workspace that helps people and caregivers manage their health information, guide them through the healthcare system, and research complex conditions by integrating all of their health data while using cutting-edge medical research search tools and frontier thinking language models.

Health Company
Live

Stoica

Building

A company built around an open-source knowledge graph visualizer, plus the tooling that lets AI agents construct and manipulate the graph itself.

Company Open source

Hippograph

Live

Knowledge graphs as structured memory. A corpus is deconstructed into the fundamentals a line of questioning actually needs.

Knowledge graphs RAG Memory
Live

Pyllments

Building

A Python framework for composing LLM apps from elements. Same chassis for a bare chatbot, a RAG system, or an API over the flow.

Python LLM apps Composable
Repo
No matching items

Writing

Technical writing, and meanderings

Build notes and benchmarks on one side. Looser thinking about technology on the other.

Technical

Notes, series, and the work itself.

May 2024

Don’t RAG on Knowledge Graphs(Or Do) Benchmarking: Combining Knowledge Graphs and Vector DBs to Answer Questions(With Sourcing) – Part Four

Building the answering segment of the QA system centering around the knowledge graph and vector db once both are populated.

knowledge-graphs rag benchmarking vector-databases
Apr 2024

Don’t RAG on Knowledge Graphs(Or Do) Benchmarking: Adding a Vector Database – Part Three

Hybridizing our Knowledge Graphs With Vector Databases

knowledge-graphs rag benchmarking
Apr 2024

Don’t RAG on Knowledge Graphs(Or Do) Benchmarking: Finally Building a Knowledge Graph – Part Two

Building a knowledge graph in Python with Claude 3 Haiku(Works for ≥ GPT 3.5 as well)

knowledge-graphs rag benchmarking
Apr 2024

Don’t RAG on Knowledge Graphs(Or Do) Benchmarking: Theory behind using an LLM to Build Knowledge Graphs – Part One

Going into the nitty gritty (theoretical) details and nuances of building knowledge graphs with Large Language Models

knowledge-graphs rag benchmarking
Apr 2024

Don’t RAG on Knowledge Graphs(Or Do): RAG, Knowledge Graphs, and Benchmarks – Part Zero

Foundational preamble to the benchmarking of knowledge graph-centric RAG flows

knowledge-graphs rag benchmarking
No matching items

All technical posts →

Thinking Out Loud

Meanderings about technology and the rest of it.

Jun 2024

LLMs, Brains, Physical Reality. Can Large Language Models really lead to superintelligence?

If LLMs are bootstrapping themselves with the product of billions of years of evolution, how would their capacity expand beyond that?

LLMs Brain AGI
No matching items

All thinking out loud posts →

Consulting

Prudent Patterns

I work with startups and established companies to get value out of their data and to build AI systems that hold up once they meet the real problem. Along the way I help stakeholders understand the current landscape well enough to decide what is worth doing and what isn’t.

If you’re thinking about a collaboration or an AI system for your business, get in touch. I’d like to hear about it.

prudentpatterns.com

 
  • © 2026 - Dmitriy Leybel