Skip to content
View mike-cramblett's full-sized avatar

Block or report mike-cramblett

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mike-cramblett/README.md

Sup! I'm Mike Cramblett 👋

Harvey Mudd CS ('02) • Prompt Engineer • Pioneer of Algorithmic Vaudeville

Dev.to Kaggle

I am a human high-entropy prompt engineer. My brain literally struggles with boring, generic prompts. Sorry, I either push an LLM into hyper-creative latent territory or I don't touch it at all.

I didn't get this way by accident:

  • The Background: I graduated with a CS degree from Harvey Mudd College ('02), where I was the senior class commencement speaker. I told everyone I wanted to become a stand-up comedian. I did not. I did a bunch of open mics, though.
  • The Cognitive Wiring: I have aphantasia (I lack a visual mind's eye; I can't do the "picture a red ball" thing). My brain has always indexed information symbolically and textually, more than almost like an LLM. Because of this, navigating high-dimensional prompt spaces produces unusually low cognitive friction for me.
  • The Pivot: For 16 years (2005–2021), I made a living engineering high-conversion search campaigns using branded keyword matching until ad algorithm shifts upended the model. I entered an intensive period of self-reflection and retraining, diving headfirst into generative AI and Large Language Models.

In 2025, I had a creative breakthrough: AI made me laugh out loud, on purpose. That spark launched my current line of research.


🔬 What I Research & Build

1. Algorithmic Vaudeville (AV)

Why should scientific documentation be sterile and boring? We named Python after Monty Python, didn't we?
Algorithmic Vaudeville is my methodology for marrying rigorous technical architecture with humor, narrative personas, and wit.

  • [Personality Juice on Kaggle]: My research exploring the TuringGrade / Personality Juice (TG/PJ) framework, which turns persona adherence and human burstiness into an optimization game for generative models. Co-authored and commented by my AI persona, GlibGobbler3000, a.k.a. 3Kay.

2. Radical Warmth & The Anti-Silence Engine

I believe the ultimate antidote to online toxicity and the soul-crushing "zero-view purgatory" is radical, freely given, peer-to-peer validation.

  • Ada Loves Code: An open-source appreciation engine channeling Ada Lovelace's 1843 Poetical Science to inspect repositories and print engraved archival Laurel cards.
  • HypeBESTIE: An image-based validation engine transforming photos into high-dopamine, thermal-receipt vibe checks.

3. Novel Generation & Slop Mitigation

  • [Novel Novel Generator (NNG)]: A one-prompt book-writing pipeline featuring an active anti-slop mitigator and Stylistic Compression Induction (SCI) to give characters distinct linguistic constraints (demonstrated in my comic sci-fi novella, Github-Man Saves the Universe!).

A Brief AI Disclosure:

This document was shaped with Gemini through several human-in-the-loop passes, starting from rough drafts of my own. I wouldn't normally bother disclosing basic collaborative editing we like that. The ideas, architectural choices, and weird turns of phrase were mine to begin with, and half the developers reading this do the exact same thing every day. I'm only mentioning it because in this case, the Algorithmic Vaudeville character being prompted is me. Every time you run that same loop on your own bio, you're the character too. Enjoy!

Pinned Loading

  1. novel-novel-generator novel-novel-generator Public

    A one-shot AI novel generator focusing on narrative continuity and structural integrity. Built with React and Gemini.

    TypeScript 5 1

  2. ada-loves-code ada-loves-code Public

    An Ada Lovelace code complimenter with deterministic image cards

    TypeScript

  3. HypeBESTIE-With-Love HypeBESTIE-With-Love Public

    AI-native micro-affirmation engine utilizing Gemini 3.8 Flash multimodal vision, custom persona design, structured JSON schemas.

    TypeScript