David An

In this article, we will be fine tuning the Llama 3.2B model with Unsloth on the Spider 1.0 SQL dataset.

Short-term gratification has taken over the world.

You have a brand new shiny GPU and want to start experimenting with it by running some sample experiments in Kubernetes, but how would you start that.

I want to preface that any opinions expressed here are of my own and not representative of any organization or my employers.

You ever just wonder how large models serve at scale?

You ever wonder what goes on when you ask ChatGPT a question and how that is served?

Lately, all we have heard about is tokenization and embeddings and the role they play in the greater LLM and AI ecosystem.

In a continuation of the previous article, we explore the implementation of these different examples.

Kubernetes is now widely used for managing containerized applications.

In today’s day and age, organizations have more data than ever.

Flights start blending together one after one.

It’s been 2.5 months since I graduated, and approximately 3-ish months since the last computer or mathematical related post.

A long overdue set of writeups (solutions) for UIUCTF23.

UPDATE: This article was covered in the Fortune Magazine article “Snap’s ‘My AI’ chatbot tells users it doesn’t know their location.

The streets are empty with the streetlights being the only company in the morning.

*This post is pending for SIGBOVIK 0x7e7* For some background information, SIGBOVIK is a yearly scientific conference hosted by Carnegie Mellon University where participants are able to flex their lesser-known research muscles.

So, how does it begin?

At the beginning of any data analytics/data science project, the most usual case is utilizing Pandas to load the data into a object called a DataFrame and perform preprocessing tasks on it.

This is a long overdue writeup with open source learning and Summer of Side Projects 2022 “Go do a project.”

This article assumes the reader has basic knowledge of calculus and analysis.

Some introductions into harmonic functions.

“$7.13 is your total sir.”

Libraries store the energy that fuels the imagination.

All of this comes together for a Fake Detection Network, or FaDe-Net.

Building an NLP-Powered Repository for Cyber Risk Literature [Poster] David An, Linfeng Zheng, Zhiyu (Frank) Quan Abstract With the large and growing body of cyber risk literature, we see three major challenges faced by the actuarial research community: t…

Scraping Web Data for Academic Literature Done in collaboration with Dr.

FaDe-Net [Writeup] David An - AP Research Project Abstract The rapid development of social media and online news outlets has accelerated the spread of fake news across the internet.