art
Großstadtdschungel (2018)
Acryl auf Leinwand (120x80cm)
Bugatti (2018)
Acryl auf Leinwand (100x80cm)
blog
- Developing a web-browsing AI agent from scratch.
- Tackling difficult problems
- Learning AI Agents
- Explainable AI: Understanding LIME for Image Data
- Explainable AI: Understanding LIME for Tabular Data
- Bayesian Regression (3/3) - Python Implementation
- Bayesian Regression (2/3) - Likelihood
- Bayesian Regression (1/3) - Model Setup
- Python Example of Gradient Descent for f(x,y)=x²+y²
- Deriving the Gradient of the Cross Entropy Loss Function
- How to generate new ideas
- The Problem with Net Revenue Retention
books
- The Story of More
- Attached by A. Levine & R. Heller
- Levels of Life by Julian Barnes
- The Book You Wish Your Parents Had Read (and Your Children Will be Glad That You Did) by Philippa Perry
misc
programming
🥇🥇🥇#1 Winner of the AI+Expo Hackathon Anthropic Track in Washington, D.C.
Built an AI-based system for building disease monitoring dashboards. I specifically made a Data Engineer AI agent that is able to search, validate, clean data for any disease - think Deep Research but for datasets.
Built a web-browsing AI agent that can book tennis courts for you. You can call it via a Telegram chat bot. Leveraged strengths of multiple LLMs (Claude, GPT, Gemini). Used Playwright for browser interaction. The agent navigates websites to perform actions like checking availability of courts and making bookings. Used self-reflection and other tricks to maximize reliability.
A tool to turn your Kindle's highlights and notes into markdown formatted documents. I use it to automate the writing of book summaries.
A university NLP project in which I programmed a recurrent neural network (LSTM) that generates customer reviews given a desired product category. I used publicly available Amazon customer review datasets.