Skip to content
View thecraigd's full-sized avatar

Block or report thecraigd

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.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

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

Report abuse
thecraigd/README.md

CraigDoesData

Senior Data Analyst | AI Safety Researcher | Tech Enthusiast


I am a Senior Data Analyst and Independent AI Researcher based in Berlin, Germany.

With over 10 years of experience (most recently at Klarna), I specialize in the intersection of rigorous statistical methodology, data engineering, and machine learning. My current focus is on AI Safety, specifically investigating alignment failures in Large Language Models (LLMs) and building robust evaluation pipelines.

I am currently seeking opportunities that blend my background in data analytics at scale with my passion for AI research and engineering.


🔬 Latest Research: AI Safety & Alignment

The Devil in the Details: Emergent Misalignment, Format and Coherence in Open-Weights LLMs
Craig Dickson (2025)

I conducted an independent study replicating and extending findings on "emergent misalignment" across 9 modern open-weights models (Gemma 3, Qwen 3). The research identified a critical vulnerability where JSON output constraints double misalignment rates compared to natural language.

Resource Description
📄 Read the Paper (arXiv) Full methodology, statistical analysis, and findings.
💻 View the Code End-to-end pipeline: LoRA fine-tuning (unsloth), vLLM inference, and LLM-as-judge evaluation.
📊 HuggingFace Dataset The full dataset of over 50,000 model responses generated during the study.
✍️ Blog: The JSON Trap An accessible breakdown of why structured outputs make models less safe.

🛠 Tech Stack

Data Engineering & Analytics
SQL Python dbt Airflow Tableau AWS

AI & Machine Learning
PyTorch HuggingFace vLLM Google Colab Pandas


🚀 Featured Projects

Alexa-to-Gemini Bridge
Python, Flask, Google Gemini API
A middleware application that upgrades legacy Amazon Alexa hardware by routing queries through Google's Gemini 2.0 Flash Thinking model. This enables complex reasoning and natural conversation on older smart speakers.

AI-powered RPG Game - The Aurora Nexus
Streamlit, OpenAI API
An immersive text-based role-playing game powered by LLMs. It features dynamic storytelling where the AI acts as the Dungeon Master, adapting the narrative in real-time based on user choices.

Universal AI Chatbot Interface
Streamlit, Multi-Provider API Integration
A versatile chat platform supporting hot-swappable providers including Google Gemini 2.0 Flash Thinking, Deepseek (r1 & Chat), OpenAI GPT-4o, and Anthropic Claude.

Guitar Buddy
JavaScript, PWA
GuitarBuddy is a mobile-optimized progressive web application designed to help musicians with essential tools for practice, learning, and performance.


Don't hesitate to drop me a line if you have any questions about my research, or want to discuss working together.

Popular repositories Loading

  1. Python_SQL Python_SQL Public

    A notebook working with Python and MySQL, using MySQL Connector.

    Jupyter Notebook 137 65

  2. SQL_School_Tutorial SQL_School_Tutorial Public

    This repo contains the associated files for my tutorial series about MySQL at Towards Data Science

    48 17

  3. Avocado-Prices Avocado-Prices Public

    Mapping Avocado Prices in the US using Python - GeoPandas, GeoPy and Matplotlib

    Jupyter Notebook 5 5

  4. California_Housing California_Housing Public

    Web Application to make predictions using an ML model.

    Python 2

  5. thecraigd thecraigd Public

    2

  6. BokehAvocado BokehAvocado Public

    Creating an interactive visualisation of avocado price data in the United States

    Jupyter Notebook 1 6