marcosomma

marcosomma @marcosomma

About: I'm Marco Somma, a cognitive systems architect, technologist, and builder of modular reasoning tools. I’m the creator of OrKa – an open framework for orchestrating explainable AI agents.

Location:
Barcelona, Spain
Joined:
Apr 18, 2025

marcosomma
articles - 56 total

🧠I Built a Support Triage Module to Prove OrKa’s Plugin Agents

A branch-only experiment that stress-tests custom agent registration, trust boundaries, and...

Learn More 3 0Jan 10

🧠Impostor Syndrome Workflow.

I Built a Multiagent Workflow to Understand My Impostor Syndrome A dark, dry, self-deprecating field...

Learn More 13 0Jan 3

How to Design Two Practical Orchestration Loops for LLM Agents

Building a useful AI assistant is no longer about a single clever prompt. Once you have tools,...

Learn More 8 0Dec 8 '25

Binary weighted evaluations...how to

Evaluating LLM agents is messy. You cannot rely on perfect determinism, you cannot just assert...

Learn More 12 2Dec 7 '25

🧠LLMs As Sensors

Why OrKa 0.9.10 Wraps GenAI Inside Deterministic Systems I will start bluntly. I like generative...

Learn More 5 0Dec 6 '25

🧠Maybe I Just Do Not Get It!

I have been working with AI for a while now. Not as a tourist. Not as a weekend hacker. Deep in it....

Learn More 49 17Dec 2 '25

🧠Veterinary AI Workflow with OrKa: Specialist Orchestration, Structured JSON Input, and Observable Reasoning

Introduction Modern veterinary medicine often deals with complex clinical cases that...

Learn More 10 0Nov 29 '25

Reasoning In The Wild: How I Actually Think About Cognition

Cognition, for me, did not start in a lab or a code editor. It started in a field, in the mud,...

Learn More 5 0Nov 25 '25

Don Quixote Of Orchestration: Building For Problems No One Sees Yet

I think I have a strange "talent" that is both a gift and a curse. I tend to see certain problems...

Learn More 14 3Nov 23 '25

🧠Deterministic scoring for messy AI agent graphs: what I learned building OrKa v0.9.6

Over the past 8 months I have been quietly building my own cognition layer for AI systems. Not a...

Learn More 7 6Nov 15 '25

🧠GraphScout plus Plan Validator: from route discovery to gated execution

TLDR: GraphScout explores candidate reasoning paths, Plan Validator grades those plans across...

Learn More 3 0Oct 27 '25

Six Months Building Open Source: What I Learned, What I Wish I Knew, What I Know Now

OrKa is not a chatbot kit. It is a modular cognitive architecture that executes graphs of purpose...

Learn More 10 3Oct 26 '25

🧠OrKa docs grew up: a YAML-first reference for Agents, Nodes, and Tools

I rewrote a big slice of OrKa’s documentation after some painfully honest feedback. The summary of...

Learn More 15 2Oct 17 '25

🧠Introducing OrKa Cloud API

When One AI Agent Isn't Enough Imagine you're building a research assistant. You ask your...

Learn More 9 0Oct 14 '25

🧠Why OrKa-Reasoning: what orcas can teach us about building smart agent teams 🐋

TL;DR Orcas live in tight family groups, speak with recognizable accents, learn skills from elders,...

Learn More 9 0Oct 5 '25

OrKa 0.9.4: cleaner logs, full GraphScout paths, ISO timestamps

TLDR: final agent is always logged, GraphScout multi hop paths execute completely, timestamps are ISO...

Learn More 2 0Oct 5 '25

🧠Building AI Beyond the Hype

Introduction Believing in a theory that runs against the dominant paradigm is never easy....

Learn More 7 0Sep 25 '25

🧠GraphScout: Self-Discovering Paths in OrKA

Introduction When I first started building OrKA-reasoning, I quickly realized a recurring...

Learn More 9 1Sep 22 '25

🧠Minsky’s six memory types as Orka preset memory.

I hate cargo-cult “memory.” Shovel everything into a vector DB and call it cognition. No. If you want...

Learn More 13 2Sep 18 '25

🧠Loop of Truth: From Loose Tricks to Structured Reasoning

The myth of "new" ideas AI research has a short memory. Every few months, we get a new buzzword:...

Learn More 5 0Sep 7 '25

🧠From 77% to 92%: How Orka-Reasoning Turns GPT-oss:20B Into a Math Reasoning Powerhouse

Local Model, Big Results Everyone assumes you need GPT‑4 or Claude 3 to hit...

Learn More 11 2Sep 5 '25

💻The HTML of AI: Why Prompting Isn't Enough!

🚀 In the late 90s I felt like a wizard. I had discovered HTML. I could open Notepad, type a few...

Learn More 20 3Sep 2 '25

🧠OrKA-Reasoning: How Workflow Execution Really Works

AI today is too often presented as a black box: data goes in, "magic" happens, and an answer comes...

Learn More 4 2Aug 29 '25

💻MCP: Protocol as Product, Hype as Standard

“Reflection MCP is a f#@$%ng protocol. Why people get so crazy about it? How much is hype and how...

Learn More 5 3Aug 25 '25

🧠Can AI Learn to Care?

How OrKa’s Maternal-Inspired Workflows Echo Geoffrey Hinton’s Call for "Maternal Instinct" in...

Learn More 14 15Aug 16 '25

🧠OrKa onboarding: example suite, guided tour, and trace replay

Engineers do not need more hype. We need runnable flows, visible state, and contracts that survive...

Learn More 17 0Aug 10 '25

🧠Building neuro-symbolic AI Alone... Help is welcome

Hi DEV, I'm Marco. Four months ago I started building OrKa, a modular, YAML-defined orchestration...

Learn More 11 4Aug 7 '25

🧠 I Didn't Know Where It Was Going. I Just Kept Going.

I didn’t build orka-reasoning with a final goal in mind. There was no pitch deck. No spec. No clean...

Learn More 16 3Jul 31 '25

💻The Flat World of AI Apps

Two years into the greatest hype wave in tech history, and what do we have to show for...

Learn More 9 3Jul 29 '25

🧠 Local, Trustable and Explainable AI. Not a dream!

Look At real inspectable Memory with TTL orka-reasoning pypi Design your...

Learn More 8 0Jul 25 '25