Bartosz Lenart
My interests
- AI & Machine Learning
- Brain-computer interfaces
- Robotics & Automation
- Neuroscience & Cognition
- Quantum computing
- Physics & Medicine
What I build
Focus: AI systems, multi agent workflows, and regulation aware stacks for domains where failure is costly.
End to end: systems, product architecture, and UI/UX.
Interface, deployment, and maintainability get the same attention as the internals.
The standard: measurable behavior, traceable decisions, and code that survives handoff.
Technology Stack:
Python · FastAPI · LangGraph · Pydantic · GenAI · ML · React · Docker · SQL · Cloud
- System evaluation & verification: benchmarks, regression catches, and explainable reasoning.
- Medical AI & FHIR native tooling for clinical workflows with auditable state.
- Financial crime prevention: AML and anti fraud pipelines with traceable decisions.
- Multi agent orchestration: typed contracts, observable handoffs, bounded tool use.
- Legacy modernization: data pipelines and migration paths for AI adoption without downtime.
BLOG.CKPT
LATEST
Bonfires in the Dark: Ritual, Science, and AI. The Interfaces We Live By.
Old ritual was not bad science but a coordination layer running under hard constraints. A look at how ritual, science, and AI each carry the work of building a world model and binding a community.
TOP.READS
The Mirror Has No Face: Why AI Only Sounds Conscious When You Ask It To
LLMs produce self-awareness language only when prompted. Empirical evidence shows zero unprompted psychological language in 100,000 neutral prompts, context-dependent register switching, not emergent consciousness.
The Event Horizon of Thought: Where Attention Meets Architecture
Exploring the hard limits of attention - whether human working memory or AI context windows - as fundamental structures to be navigated. A framework for designing workflows that respect the physics of thought.
AI as Cognitive Prosthetic: You are the art, AI is an interface, the result is value.
Moving from answer generation to insight extraction. Agentic systems should augment human thought. They function as high-performance 'cyberpunk chrome' designed to extract and amplify the wisdom already present in your mind.
The Map Without a Vision: Your Brain Navigates a World You Can't Picture
Much of brain computation stays below conscious access. Only a small subset of competing states ever wins broadcast. A family of conditions shares one fault line where the map and the report come apart. That split is also the architecture we are building agents on.
Yggdrasil: Parallel AI Reasoning Architecture
A single tree grows in one direction. Yggdrasil doesn't just explore parallel paths. It detects when insights from one cognitive realm unlock breakthroughs in another, then builds the bridge.
The Geometry of Compression: Why Angles Survive What Raw Storage Loses
Angles survive what raw storage loses. A visual essay on Moroccan zellige, transformer KV-cache compression, and TurboQuant. Good compression keeps the pattern that still lets a system work after deletion.
The Neural Bifrost: Bridging Deep Brain Signals and the Surface of Consciousness
A 2026 paper put electrodes inside ten human brains while patients tried to see a near-invisible flicker. The result is a high-resolution view of how conscious access is gated, and the architecture is a bridge, not a destination.
NEWSLETTER
Essays and field notes on AI, cognition, consciousness, and the architectures that shape how we think, build, and coordinate.
PROJ.DIR
Yggdrasil Agent
Agentic reasoning platform that solves complex tasks by exploring parallel solution paths. Ships with four interfaces (REST API, MCP, CLI, Chainlit UI), plugs into any LLM vendor, and scores 72% on the GAIA benchmark. Built to integrate into real workflows, not just demo well.
FHIR Flame
FhirFlame is a modular, Dockerized medical AI toolkit for processing healthcare data with multi-agent workflows. It features FHIR R4/R5 compliance, NLP entity extraction, DICOM support, local/cloud deployment (GPU-optimized), and secure API access.
HelloFibro
Chronic pain management app designed for fibromyalgia, combining supportive conversation, medication routines, and appointment preparation in one calm interface. Runs a MedGemma medical model pathway behind red-flag safety checks, with a 21-language interface and in-memory session isolation.
Can Run
CanRun is a game compatibility checker that works as a NVIDIA G-Assist Plugin. It instantly analyzes your PC's hardware against game requirements, predicts performance, and provides upgrade recommendations-all through simple voice commands or API calls.
Turning complex reasoning into reliable AI systems.
I help teams ship explainable AI systems, exploring agent workflows, and financial crime detection platforms that hold up under real operational and compliance pressure.
- 15+years of embracing computer science, focused on AI-empowered systems
- 5+years designing financial crime detection for regulated environments
- PSM Ⅰcertified Professional Scrum Master
I enjoy exchanging notes with founders, researchers, engineers, clinicians, and operators working on agent reliability, medical AI, financial-crime systems, and cognitive architectures.
Bring a difficult problem, a contrarian thesis, a paper, or an ambitious idea.