Posts by Tags

ai

What is it like to do AI research inside Nike?

6 minute read

Published:

I’m feeling bittersweet because yesterday, December 19, 2025, was my last day at Nike. My six-month research internship culminated in three presentations across multiple time zones. While the details of the work will eventually show up in my MIT thesis next year, here I reflect on this unique journey of conducting AI research at a world-renowned brand.

class

Cheatsheet for Your MIT Sloan Classes

1 minute read

Published:

Hi there. If you wind up on this page, I assume you’re busy studying for your upcoming MBA Core exams at MIT Sloan.

6.8300 Final Project: Taming CLIP’s Captioning Bias: A COCO-Driven Analysis and Permutation Ensemble

20 minute read

Published:

Abstract: Vision-language models like CLIP struggle with multi-object scenes, often favoring prominent objects or those mentioned first in captions. Using real-world COCO images, we show that CLIP’s caption-matching accuracy drops from 91.23% to 87.45% when object order is reversed. To address this, we explore a post-hoc mitigation: a permutation ensemble that averages scores across all object orders, boosting robustness and recovering accuracy to 90.04%. Our findings reveal persistent order biases and offer a simple, effective strategy to improve CLIP’s reliability in complex scenes.

computer vision

6.8300 Final Project: Taming CLIP’s Captioning Bias: A COCO-Driven Analysis and Permutation Ensemble

20 minute read

Published:

Abstract: Vision-language models like CLIP struggle with multi-object scenes, often favoring prominent objects or those mentioned first in captions. Using real-world COCO images, we show that CLIP’s caption-matching accuracy drops from 91.23% to 87.45% when object order is reversed. To address this, we explore a post-hoc mitigation: a permutation ensemble that averages scores across all object orders, boosting robustness and recovering accuracy to 90.04%. Our findings reveal persistent order biases and offer a simple, effective strategy to improve CLIP’s reliability in complex scenes.

culture

Stolen Name: How Software Erases Identity

5 minute read

Published:

Names are one of the first things a system asks for, and one of the easiest ways it reveals what culture it was built for. This post looks at how seemingly harmless assumptions in software can quietly erase identity, using Chinese names as the primary case study.

grad-school

Networking Beyond Your Home Department: An MIT CS/AI Case Study

4 minute read

Published:

How do you build real connections outside your home academic program or department, especially in a place like MIT? This post shares a simple playbook for doing exactly that, using my own path into the CS/AI community as a case study. The short version: find the right announcement surfaces, show up (even when you won’t understand everything), and keep the curiosity dial set to “loud.”

internship

What is it like to do AI research inside Nike?

6 minute read

Published:

I’m feeling bittersweet because yesterday, December 19, 2025, was my last day at Nike. My six-month research internship culminated in three presentations across multiple time zones. While the details of the work will eventually show up in my MIT thesis next year, here I reflect on this unique journey of conducting AI research at a world-renowned brand.

latex

Tutorial: Use LaTeX Locally with VS Code

6 minute read

Published:

This post is a quick guide to setting up LaTeX on your own computer, so you can write papers, resumes, or dazzling “class notes” using LaTeX locally—without relying on internet connectivity. In other words, you won’t be tied to Overleaf: you can bring your laptop anywhere and keep working on your next big ideas.

mit

What is it like to do AI research inside Nike?

6 minute read

Published:

I’m feeling bittersweet because yesterday, December 19, 2025, was my last day at Nike. My six-month research internship culminated in three presentations across multiple time zones. While the details of the work will eventually show up in my MIT thesis next year, here I reflect on this unique journey of conducting AI research at a world-renowned brand.

Networking Beyond Your Home Department: An MIT CS/AI Case Study

4 minute read

Published:

How do you build real connections outside your home academic program or department, especially in a place like MIT? This post shares a simple playbook for doing exactly that, using my own path into the CS/AI community as a case study. The short version: find the right announcement surfaces, show up (even when you won’t understand everything), and keep the curiosity dial set to “loud.”

Cheatsheet for Your MIT Sloan Classes

1 minute read

Published:

Hi there. If you wind up on this page, I assume you’re busy studying for your upcoming MBA Core exams at MIT Sloan.

6.8300 Final Project: Taming CLIP’s Captioning Bias: A COCO-Driven Analysis and Permutation Ensemble

20 minute read

Published:

Abstract: Vision-language models like CLIP struggle with multi-object scenes, often favoring prominent objects or those mentioned first in captions. Using real-world COCO images, we show that CLIP’s caption-matching accuracy drops from 91.23% to 87.45% when object order is reversed. To address this, we explore a post-hoc mitigation: a permutation ensemble that averages scores across all object orders, boosting robustness and recovering accuracy to 90.04%. Our findings reveal persistent order biases and offer a simple, effective strategy to improve CLIP’s reliability in complex scenes.

mitsloan

Cheatsheet for Your MIT Sloan Classes

1 minute read

Published:

Hi there. If you wind up on this page, I assume you’re busy studying for your upcoming MBA Core exams at MIT Sloan.

music

Jazz and Homesickness

9 minute read

Published:

On a cold January day in 2023 in Nashville, Tennessee, just before heading to the airport, I took a small detour to stop by Nashville Jazz Workshop. The venue was closed, but I stood outside for a moment, feeling as if a smooth piano melody were drifting through the air, quietly comforting my homesick heart. If this pairing of American jazz and my Chinese homesickness catches your attention, I invite you into a very personal story of mine, one that intertwines the life journey of an international student with a particular American jazz pianist. Happy 2026, from my home in Shenzhen, China.

networking

Networking Beyond Your Home Department: An MIT CS/AI Case Study

4 minute read

Published:

How do you build real connections outside your home academic program or department, especially in a place like MIT? This post shares a simple playbook for doing exactly that, using my own path into the CS/AI community as a case study. The short version: find the right announcement surfaces, show up (even when you won’t understand everything), and keep the curiosity dial set to “loud.”

nike

What is it like to do AI research inside Nike?

6 minute read

Published:

I’m feeling bittersweet because yesterday, December 19, 2025, was my last day at Nike. My six-month research internship culminated in three presentations across multiple time zones. While the details of the work will eventually show up in my MIT thesis next year, here I reflect on this unique journey of conducting AI research at a world-renowned brand.

personal

Jazz and Homesickness

9 minute read

Published:

On a cold January day in 2023 in Nashville, Tennessee, just before heading to the airport, I took a small detour to stop by Nashville Jazz Workshop. The venue was closed, but I stood outside for a moment, feeling as if a smooth piano melody were drifting through the air, quietly comforting my homesick heart. If this pairing of American jazz and my Chinese homesickness catches your attention, I invite you into a very personal story of mine, one that intertwines the life journey of an international student with a particular American jazz pianist. Happy 2026, from my home in Shenzhen, China.

projects

6.8300 Final Project: Taming CLIP’s Captioning Bias: A COCO-Driven Analysis and Permutation Ensemble

20 minute read

Published:

Abstract: Vision-language models like CLIP struggle with multi-object scenes, often favoring prominent objects or those mentioned first in captions. Using real-world COCO images, we show that CLIP’s caption-matching accuracy drops from 91.23% to 87.45% when object order is reversed. To address this, we explore a post-hoc mitigation: a permutation ensemble that averages scores across all object orders, boosting robustness and recovering accuracy to 90.04%. Our findings reveal persistent order biases and offer a simple, effective strategy to improve CLIP’s reliability in complex scenes.

research

What is it like to do AI research inside Nike?

6 minute read

Published:

I’m feeling bittersweet because yesterday, December 19, 2025, was my last day at Nike. My six-month research internship culminated in three presentations across multiple time zones. While the details of the work will eventually show up in my MIT thesis next year, here I reflect on this unique journey of conducting AI research at a world-renowned brand.

software

Stolen Name: How Software Erases Identity

5 minute read

Published:

Names are one of the first things a system asks for, and one of the easiest ways it reveals what culture it was built for. This post looks at how seemingly harmless assumptions in software can quietly erase identity, using Chinese names as the primary case study.

tutorial

Tutorial: Use LaTeX Locally with VS Code

6 minute read

Published:

This post is a quick guide to setting up LaTeX on your own computer, so you can write papers, resumes, or dazzling “class notes” using LaTeX locally—without relying on internet connectivity. In other words, you won’t be tied to Overleaf: you can bring your laptop anywhere and keep working on your next big ideas.

vscode

Tutorial: Use LaTeX Locally with VS Code

6 minute read

Published:

This post is a quick guide to setting up LaTeX on your own computer, so you can write papers, resumes, or dazzling “class notes” using LaTeX locally—without relying on internet connectivity. In other words, you won’t be tied to Overleaf: you can bring your laptop anywhere and keep working on your next big ideas.