About Me
Alberto Diaz Durana Mechanical engineer by training, data scientist by career, system thinker by nature.
Berlin, Germany
I design systems that help humans and AI work together without the human becoming a rubber stamp.
I started as a mechanical engineer in Colombia, writing my thesis on the biomechanics and mechanical advantage of the human locomotion system – Euler angles and long hours in the motion lab. After graduating I moved into designing hydroelectric plants and leading teams of engineers and designers. Then I came to Berlin, earned an MSc in Process, Energy, and Environmental Systems Engineering at TU Berlin – my thesis with WWF-Germany was modeling a cooling system for a server room from the thermo-economic and exergo-economic perspective, solving complex systems of equations in EES. Later I certified as a PMP and Lean Six Sigma practitioner. From there, a decade crossing boundaries: energy consulting at GETEC, BI systems at WWF, project management at Green City Solutions, co-founding a sustainability startup (Hedera Sustainable Solutions GmbH), process mining consulting at Appian, and deploying ML models for cement manufacturing at Alcemy.
I learned to code with C++ in 2001, when laptops did not exist, there was one central computer room in the university, and the reference was a book. I taught myself Python in 2013 through videos, Stack Overflow, and trial-and-error. Things before LLMs were different.
The thread through all of this is the same: modeling physical systems during my studies, certifying in project management, and a lasting fascination with philosophy taught me to see projects as systems of interacting knowledge areas, where internal and external factors evolve through phases within constraint boundaries. Philosophy shaped how I see and act on things as much as engineering did. I understand physical systems, I understand data, and I have fun connecting the two in ways that make sense to the people who have to use them.
What I do now
I created Take AI Bite, a framework for structured human-AI collaboration. It started as personal notes for managing data science projects with AI coding agents. It became 9 principles, a versioned methodology (DSM), 14 repositories, 84 features across 11 capability domains, and 20,000+ lines of documentation governing how I work with AI every day.
Every line of code in my projects reflects a deliberate decision. Sometimes I write it, sometimes I edit or correct what the AI produces, sometimes I approve a suggestion that already matches what I need. Nothing ships without my judgment. I run multiple simultaneous projects, reviewing one while another generates output. That takes comprehension and the discipline to say no.
When the documentation grew complex enough to need its own integrity checking, I built Graph Explorer, a 547-test validation tool that parses cross-references, builds knowledge graphs, detects semantic drift, and maps connections across the DSM documentation network. Using the methodology to build the tooling that validates the methodology.
The spoke projects (SQL Query Agent, DevFlow Analyzer, RAG Document Assistant) show hands-on engineering. Take AI Bite and DSM show something else: designing systems and maintaining quality across scale.
What I care about
I enjoy reading and connecting ideas. Philosophy, literature, technical papers, podcasts. I have spent years moving between technical domains, going deep enough in each to do real work before carrying what I learned into the next one. Deep reading, structured thinking, patience to understand before acting.
Words carry meaning and impact, something I learned through commercials, acting on stage, and singing in public. When I present a technical concept, people tend to listen. That has shaped every consulting engagement and every team I have led.
I care about building things that last. The methodology has an 82% backlog completion rate across 300+ items because I treat it like a product.
Why I built this
I could not find what I was looking for. Every AI workflow I encountered optimized for speed: generate more, approve faster, ship sooner. None of them answered the question I actually cared about: how does the human stay meaningfully present when the machine can produce faster than anyone can read?
So I built my own answer. Take AI Bite is a way of working with AI where the human directs, reviews, and decides, as the source of what makes the work worth doing. The idea, the plan, the vision, the spark, the judgment, the sense of when something is right. AI amplifies that, it does not replace it.
This is for people who want to build with an agent, not just hand things off to one. Who care about accumulating experience through collaboration. Who think critically about what gets produced and choose with intention what ships.
I speak Spanish (native), English (C2), German (C2), and Portuguese (B2). I hold a PMP certification and a Lean Six Sigma credential, and I have 10+ years of international experience across Colombia, Germany, Austria, and projects in 15+ countries.
If any of this resonates, I am always happy to have that conversation.
Everything else
I value nature. I do not understand it all, and that is part of what I value – how it finds order, how it prevails. I recharge with simple things: a landscape, a sunset, stars, art, music, good food, a freezing lake (thank you, Berlin), dancing, a walk in the snow, or in the sun with music or a podcast.
I am drawn to the human body as a system. My mechanical engineering thesis was on biomechanics, and I never lost the interest: how joints, tendons, ligaments, muscles, food, and energy influence how I feel and perform. I value strength training, running, biking, and riding a unicycle while juggling three balls in the air. My favorite way to rest my mind is good music and a dedicated workout.
I play guitar. I like looking at music in detail – decomposing sounds and rhythms, levels and intensities, how the instruments in a band sound connected. I like the idea that philosophy, science, creativity, and art belong closely to each other. Gilles Deleuze brought inspiration here, including the rhizome: a metaphor for non-hierarchical, non-linear, interconnected systems. See the pattern? Philosophy at work is collected in DSM_6.
Music I cannot stop listening to and to which I am deeply grateful for so much energy: Alice in Chains, Bach, Brahms, BRMC, Chopin, Deftones, Dozer, Dvorak, Low Rider, Massive Attack, Meshuggah (!), Miles Davis, Nine Inch Nails, Queens of the Stone Age, Radiohead, Red Fang, Shostakovich, Soundgarden, Stone Temple Pilots, Tchaikovsky, Thou (!), Tool, TV on the Radio (Return to Cookie Mountain – oh yes), YOB (!).
People who have influenced me in different ways: A.M. Homes, Albert Camus, bell hooks, Byung-Chul Han, Chris Cornell, Frankenstein (it counts as people, of course Mary Shelley), Fyodor Dostoevsky, Henry Miller, Josh Homme, Marshall Rosenberg, Mikhail Bulgakov, Nicolas Diaz Durana, Nietzsche, Oscar Wilde, Paul Auster (4 3 2 1), Simone de Beauvoir, Steven West, Wittgenstein, to name a few… and my friends and ones I love.