AI Engineer • Former Particle Physicist ⚛️ • Former Risk Modeler 📈

I've spent my whole career looking inside systems that would rather stay opaque — first particle collisions, then risk models, now neural networks.

The models talk to us all day. I want to be able to affect them back.

So I build tools that make neural networks less mysterious — because interpretability shouldn't stay in research papers, it belongs in production. (And in zombie games. And in mushroom generators. You'll see.)

The house rules, in every repo: each experiment ships with a placebo control, its negative results stay published, and when a finding turns out to be about my own instrument instead of the model, that page says so. Half the fun is the honest notes.

the-paper-that-remembers-itself · steeropathy · multi-shroomed-bandit · once-upon-256-dim · sample-me-softly-with-this-prompt · the-shroomifier · attack-on-applepear

(Every one of those is a real, running project. I trained several of them on my laptop CPU, because why not.)

My personal site is a chat with a tiny LLM running entirely in your browser, and every answer it generates renders as a real particle collision. Click the event below to fire your own question into the chamber:

One engine for looking inside a model, one factory for the directions it steers with, and the experiments that run on both. Click any box to open its repo.

flowchart TD
    hd["🧭 <b>hidden-directions</b><br/>turn a behavior into a vector —<br/>inject it, bake it into the weights,<br/>or catch a bake"]
    bs(["🧠 <b>brainscope</b><br/>watch a model think: an OpenAI-compatible server<br/>with a live view into the residual stream<br/>— logit lens · attention · J-lens · live steering"])
    st["🕹️ <b>steeropathy</b><br/>agents that talk through activations<br/>and J-space, never text — moods, concepts,<br/>a zombie outbreak fought by mind-reading"]
    tm["⚖️ <b>in-two-minds</b><br/>catch an agent hesitating between two tools,<br/>in its activations before it commits"]

    hd -->|"directions &<br/>baked personas"| bs
    bs -->|"hosts the model,<br/>captures activations"| st
    bs -->|"hosts the model,<br/>captures activations"| tm
    st -.->|"steers with"| hd

    click hd "https://github.com/moudrkat/hidden-directions"
    click bs "https://github.com/moudrkat/brainscope"
    click st "https://github.com/moudrkat/steeropathy"
    click tm "https://github.com/moudrkat/in-two-minds"

    classDef engine fill:#1f6feb,stroke:#1158c7,color:#ffffff;
    classDef exp fill:#8957e5,stroke:#6e40c9,color:#ffffff;
    class bs,hd engine;
    class st,tm exp;

The two blue boxes are the instrument. brainscope hosts any Hugging Face model and streams its internals to the browser; hidden-directions makes the steering vectors — and bakes them into weights, then audits for the bake. The two purple boxes are experiments run under that lens. steeropathy wires agents together through activations instead of text; in-two-minds catches an agent hesitating between tools before it commits.

Smaller, self-contained ways to look inside:

• 📜 paper-remembers — Hopfield's 1982 paper, running live: rub out any part of the page and watch it rebuild itself
• 🎬 green-room — a multi-character agent has already cast the scene before it writes a word of dialogue; see it in the activations, then recast it
• 🎭 sixteen-voices — how a tiny transformer encodes writing style, through LoRA adapters and attention heads
• 👁️ show-me-your-attention — attention maps and neuron activations over your own prompt
• 💥 detektor — the collision chamber above, open source (SmolLM2 in your browser, no server)
• 🖼️ jepa-demo — I-JEPA & V-JEPA 2 hands-on, no GPU needed, with a visual deep-dive article
• 🍄 Mushroom-generator — a VAE growing mushrooms, with latent-space walks and the decoder taken apart layer by layer
• 🍎 Applepear — apples vs pears in a tiny CNN, activations and grad-CAM included
• ⚙️ Minimize_me — race TensorFlow optimizers across loss landscapes

• 🎨 personal-rembrandt — you can't build a personal brand, so build a personal Rembrandt: paste your bio, GPT-2 reads it in your browser, and its activations repaint his 1659 self-portrait — the brand-safe words sink into shadow. Teaches you very little about the model and a lot about Rembrandt
• 🛏️ go-to-damn-bed — a Claude Code skill that sends you to bed like a mom sends naughty children: it saves your work into TOMORROW.md, then counts to three. It never says what happens at three
• 👑 KingOfDiamonds — the King of Diamonds game from Alice in Borderland, played by LLMs in character, recursive strategic thinking and all
• 🗨️ paralel-discordverse — your company's Discord gets a parallel universe, populated entirely by fictional colleagues
• 🧮 least-squares-method — code archaeology: a printed Pascal listing, photographed page by page and revived on Turbo Pascal 5.5

None of it is perfect. That's kind of delightful.