Build sharper paths through AI, knowledge & complex systems. An independent systems studio at the edge of AI and the built world — shipping from the terminal, not hype.
The latest articles
Long-form essays on AI engineering, knowledge management, agentic systems, and the built world. Each piece is researched, source-cited, and built to be useful — not a hot take, not a thread-bait.
- Graphiti: the temporal knowledge graph that gives agents a memory of time · 2026-09-09 · 14 min
Most knowledge graphs are a snapshot frozen at index time. Graphiti is built for the opposite: facts that change, get superseded, and need to be queried as they were true. Here is how temporal context graphs work, why they beat static GraphRAG for agents, and a working Docker Compose stack to run one tonight.
- Local AI is just a file and a program. Here is how to run it yourself · 2026-09-08 · 14 min
Running AI locally sounds like it needs a PhD and a $10,000 computer. It does not. Local AI is a model file on your disk and a program that runs it — no cloud, no API keys, no subscription. Here are the real building blocks and the four ways to run a model on your own machine, from a clickable app to pure code.
- LangChain: the agent framework that killed provider lock-in · 2026-09-08 · 18 min
LangChain 1.0 collapsed a sprawling ecosystem into one import path and made create_agent the standard way to build agents. This is the full tour — from a weather tool to structured output, memory, RAG, and the middleware system that lets you swap prompts and models at runtime — updated through the latest release, v1.4.0, which brings MCP support into the core package, and set in the modern three-layer stack of Deep Agents, LangChain, and LangGraph.
- How diffusion models generate images: the math, the architecture, the samplers · 2026-09-02 · 22 min
A diffusion model does not "draw" an image — it learns to reverse a process that turns an image into noise, then walks that reverse path from pure noise back to a picture. This is the full engineering tour: the DDPM objective, score matching and the SDE/ODE view, the U-Net-to-DiT architecture shift, the samplers that make it fast, and the conditioning stack that makes it obey a prompt.
- The map is not a picture. It's a database: the BIM lesson geospatial must borrow before AI gives up on maps · 2026-08-31 · 18 min
Construction and cartography independently invented the same idea: a model as a database of objects with properties, relationships, and measurements, where geometry is one view, not the definition. Biology has a name for that pattern — convergent evolution. This article traces the two lineages, verifies the convergence in the standards themselves, and shows why AI is the selection pressure that makes the commitment urgent.
- The Architect's New Instrument: How AI Agents Are Rewiring the Entire Design-to-Operations Workflow · 2026-08-31 · 45 min
AI agents are not a faster pencil — they are a new agent of production. From site analysis and generative ideation to drafting, fully functional BIM, and facility management, this is how the data generated at every stage compounds into value, and why the architect who does not integrate AI is forfeiting the data dividend.
Tutorials that ship
Hands-on, real, source-verified lessons. A fresh tutorial lands here every day — a Linux command in the morning, a pandas method an hour later — plus deep-dives into the stack that runs this site. Each one comes with a working example, a CLI-reference cheat table, and the honest pros and cons of using it in production.
- wget — download files and crawl sites from the terminal · 2026-09-11 · 5 min
The command-line downloader that pulls files, mirrors directories, and survives dropped connections — no browser required.
- pandas Series() — the one-dimensional column everything else is built from · 2026-09-11 · 4 min
The 1-D labeled array behind every DataFrame column: build it from dicts and lists, watch indexes align, and learn why a column is a Series but a row is too.
- curl — transfer data with URLs · 2026-09-10 · 6 min
Fetch web pages, download files, and talk to APIs from the terminal. The universal HTTP client.
- pandas DataFrame() — build a table from dicts, lists, arrays, and Series · 2026-09-10 · 5 min
The pd.DataFrame() constructor is the front door to pandas: turn records, column dicts, arrays, and Series into one table — with index, columns, dtype, copy.
- ping — test network reachability and latency · 2026-09-09 · 5 min
Send ICMP echo requests to a host and measure round-trip time. The first tool you reach for when the network misbehaves.
- LangChain 1.0 Masterclass — agents, tools, memory, RAG & middleware · 2026-09-08 · 14 min
A hands-on tour of LangChain 1.0 (released Oct 22, 2025): create_agent, init_chat_model, streaming, structured output, context, InMemorySaver memory, multimodal input, a RAG retriever tool, and the middleware system that makes create_agent the standard way to build agents.
- Mermaid — diagrams as code · 2026-09-08 · 14 min
Write flowcharts, sequence diagrams, Gantt charts and more as plain text that renders to SVG — and lives in git alongside your code.
- Harness engineering — the craft of building around a model · 2026-09-08 · 13 min
The model is the engine, the harness is the car. Context, tools, automations, evals — the four levers that turn a raw LLM into something that actually does your job.
- ss — inspect sockets and network connections · 2026-09-08 · 6 min
The modern replacement for netstat. List listening ports, established connections, and socket stats in milliseconds.
- pandas filter() — keep columns and rows by name, substring, or regex · 2026-09-08 · 4 min
df.filter() picks labels, not values: keep columns whose names match a list, a substring, or a regex — plus row filtering with axis=0.
- ip — inspect and configure network interfaces · 2026-09-07 · 6 min
The modern replacement for ifconfig. Show addresses, routes, links, and neighbors in one tool.
- pandas sample() — draw random rows for tests, demos and quick QA · 2026-09-07 · 4 min
Random sampling in one call: n rows, a fraction, per-row weights, with or without replacement. Reproducible via random_state — the daily tool for demos and holdouts.
- free — report real memory usage & swap · 2026-09-06 · 5 min
See how much RAM is actually in use — and how much is just being held as cache that Linux will hand back.
- pandas between() — one call for range filters, instead of chained comparisons · 2026-09-06 · 4 min
Range filters in one readable call: s.between(20, 40) replaces the (s >= 20) & (s <= 40) chain, and even works on dates and per-row boundaries.
Tools that ship
Single-purpose agent tools, composable like unix pipes. Each one is published to npm, versioned honestly, and built to be run from a terminal — no dashboards, no marketing sites, just throughput.
- kmailai v0.2.1 · stable
the Kmail AI coding agent, in your terminal
- kmail-memo-agent v1.0.0 · stable
unified memory + semantic grep for AI agents
- kmail-save-tokens v0.1.3 · stable
translate coding-agent prompts into token-efficient instructions
- eu-ai-act v1.0.0 · stable
navigate, understand and apply the EU AI Act from your terminal
- bim-execution-plan v0.1.0 · stable
author, version and validate ISO 19650 BIM Execution Plans
- ascii-studio v3.0.0 · stable
turn any image into ASCII art, right in your browser
- mermaid-canvas v1.0.0 · stable
write Mermaid code, watch the graph render live
Knowledge management, the discipline
Turn what an organization knows into what it ships. A field guide to knowledge management — its concepts, its process, its methods, and the technology that powers it today. Each topic lives on its own page so you can go as deep as you want, and five interactive tools (SECI spiral, knowledge tree, knowledge audit, tacit-knowledge wizard, and a RAG pipeline) let you put the theory to work.
- Core concepts
Tacit, explicit & implicit knowledge, the DIKW staircase, SECI, ba, absorptive capacity, communities of practice — the foundational vocabulary of the discipline, with citations and real examples.
- The 5-phase process
Identify → Collect → Secure → Transfer → Use. The cycle that turns scattered knowledge into a working organizational capability, with pitfalls and best practices at every step.
- Methods & techniques
After-action reviews, knowledge audits, peer assists, Knowledge Cafés, storytelling — the proven techniques for moving tacit know-how into organizational memory.
- Technology
Knowledge graphs, vector search, RAG, docs-as-code, semantic search, agent memory, MCP, GraphRAG — the modern stack that powers knowledge systems in 2026.
- The knowledge tree
Watch knowledge grow — from tacit roots to a trunk of record, branches of practice, leaves of outcomes. A visual model of how knowledge scales inside an organization.
- A practical KM method
A working, industry-proven method for keeping knowledge alive — built around the knowledge document and a four-phase process you can run Monday morning.
→ open the field guide · → glossary
One engineer, three pillars
Mohammed Kmail trained as an architect at TU Wien and has worked on real places — hotels in the Emirates, the Grand Hyatt, the full arc of planning, design, and execution — before the drawing board taught him that buildings are also data. From there, parametric design in 2014, BIM and the BIM manager role, and the move into data and programming. Today he works at the public sector on the class of agentic AI development: systems that don’t just learn, but plan, act, and build.
The three pillars of this site are knowledge management (the discipline of making what an organization knows work harder than the people who happen to hold it right now), AI expertise (machine learning and agentic AI engineered end-to-end, with the arc now pointed at systems that think), and architectural design & BIM (the built world origin that grounds the rest — buildings were the first problems to teach him systems thinking).
Everything on this site is shipped from a terminal, not from a marketing site. Tutorials, tools, a full knowledge management field guide, and the long-form essays that pin the work to the people who care about it. The build, the deploy, and the daily lessons are open — read them, fork them, and ship from the terminal too.