Here's me and the bar that I built.
RANDOM COCKTAIL
Click to generate a cocktail
ABOUT ME
I'm an Amazon scientist, and I love to build things. My background is in neuroscience, which means I've spent way too much time thinking about brains both biological and artificial.
My interests are LLMs, Brains, AI Alignment, and how technology will shape society. I also like woodworking, friends, and beer.
More professional stuff on LinkedIn.
BLOG POSTS
Convert your Medium Weekly Digest to a Podcast with NotebookLM
Jan 17, 2024
I'm a big Hard Fork listener, and wish there was a way to get more tech stuff crammed into podcast form. Nothing beats professional…
AI, Director's Cut: You're the Star in the Movie of Tomorrow
Jun 24, 2023
My mother-in-law has a peculiar fondness for JibJab. The eccentric yet entertaining platform takes public domain music videos, superimposes…
The Creator's Cut: Advocating for Paid Training Data in AI
May 12, 2023
Let's say I'm a data scientist, a futurist, a technophile. Let's say I even get a little too enthusiastic when it comes to AI and…
Maximizing Accuracy for Rare Occurrences: A Zero-Inflated MSE Approach to Model Training
Apr 4, 2023
Training a model when most predictions should be zero.
Failing Upwards: my journey from a PhD in Neuroscience to a Data Scientist at Amazon
Mar 18, 2023
Embracing uncertainty and finding a new career path in data science
Use BeautifulSoup and WordCloud to visualize the most frequent words in Firefly
Feb 22, 2022
I love firefly. Does it hold up? I don't know. But its script is published online, so we'll use it as an example to display textual data in…
Common metrics for evaluating natural language processing (NLP) models
Feb 18, 2022
You can't train a good model if you don't have the right evaluation metric, and you can't explain your model if you don't understand the…
Explain your text classification model with Lime
Feb 4, 2022
The more complex a model, the more of a black box it can be — that is, it can get pretty tricky figuring out just why your predictions…
Techniques to represent text for NLP
Jan 11, 2022
We can't just take a paragraph and feed it directly into a mathematical algorithm; the characters, words, or phrases need to be converted…
Create an automatically expanding TextEditor with a Placeholder in SwiftUI
Nov 27, 2021
I've banged my head against the desk for a while about this, and there are quite a few buggy hacks about this floating around the web, but…
Create placeholder data with GeoPoint and Date values, and upload it to Firebase Firestore
Jul 26, 2021
At some point during the development process, you may find yourself wishing you could preview what your app would look like if it had a lot…
Auto-Scrolling with ScrollViewReader in SwiftUI
Jul 22, 2021
You're building…maybe…a social media app. Or a list app. Or whatever. You have a list of items, and a way of adding items to the list…
Know when all the asynchronous functions in a loop are finished using DispatchGroup in Swift
Jul 21, 2021
Here's a pretty common scenario: You need to upload a bunch of items to a server, one by one, and get the result when it finishes. Each…
Zero to Pipeline: Beginner's Guide to Building a Scikit-learn Pipeline
March 2021
A step-by-step tutorial on building a machine learning pipeline to predict Titanic survival. Perfect for those just getting started with scikit-learn and data science workflows.