Other works
Brief Me
Personal Project
“Brief Me” is an online tool that takes the most recent news articles from a news API that match your personal preferences to use the ChatGPT API to provide AI generated summaries of the articles which can also be dictated in various languages. This tool aims to save you time when consuming the news.
Background
Science I have the habit of reading the news every day, so I’ve long struggled with two critical pain points in news consumption.
Large Amount of Information
From global headlines to niche topics—creates information overload; sorting through dozens of full-length articles to find what matters takes hours, which most people can’t spare.
Language Barriers
Many high-quality articles are only available in specific languages, and even with translation tools, converting entire articles into a readable, concise format remains tedious.
These frustrations became the core motivation for building "Brief Me"—a tool that doesn’t just deliver news, but delivers it in a way that fits how people actually want to consume it:
Quickly Familiar Language Personalized
Competitive Product Analysis
I analyzed 6 leading news tools and AI summarization services to identify feature gaps and position "Brief Me" competitively:
This comparison confirms that no existing tool combines personalized curation, high-quality AI summarization, and multi-language audio delivery—the three pillars of "Brief Me."
Design Goal
The design goals for "Brief Me" are rooted in solving the user needs identified in research—with a sharp focus on simplicity, efficiency, and accessibility. Every design decision was guided by three core objectives:
Minimize News Consumption Time
Ensure users can grasp key news insights in 1–2 minutes per topic, compared to 5–10 minutes reading full articles. This means prioritizing concise AI summaries and eliminating friction.
Personalization Without Complexity
Let users set preferences in 60 seconds or less, with the tool automatically refining recommendations over time. No technical setup or manual curation required.
Break Language Barriers
Make summaries accessible to non-native speakers by offering both written translations and audio dictation in 12+ languages. The dictation feature should also serve users on-the-go who can’t read text.
Design Outcome
Prototype
Interface
Video Capturings
Hexapod Leg Development
2023 May - 2024 March
Personal Project
I’ve had a deep passion for robotics since 4th grade, starting with VEX IQ—where I competed for 4 years, even winning the VEX IQ Worlds 2020-2021—and later moving to VEX VRC in 9th grade. While VEX challenges honed my critical thinking and design skills, they limited my creativity with pre-made parts and strict rules. To expand my skill set, I wanted to learn 3D modeling and inverse kinematics. This project—designing and building a basic hexapod leg—was the perfect opportunity: it let me practice 3D modeling/printing freely and apply inverse kinematics in a manageable, low-risk setup. I also aimed to use these new skills to improve my future VEX projects, like digital prototyping, and lay groundwork for learning engineering and design later.
Design Process
I followed a structured timeline from September to April, breaking the project into clear, actionable steps:
First, I researched 3D modeling software. I initially chose OnShape on September 18 but switched to Fusion360 on October 9 for better compatibility with my design needs. By October 1, I’d started drafting the hexapod leg components—splitting the design into subsystems to stay focused, a method I’d used in VEX robotics.
Next, I researched and ordered motors by October 15, ensuring they were compatible with Arduino. By November 30, I’d fully modeled all parts in Fusion360, 3D printed them, and assembled the mechanical structure—prepping for wiring. I finished wiring the Arduino components early and tested each servo on December 2 to fix any wiring or design issues.
For the coding phase, I spent December researching inverse kinematics to understand how to calculate joint rotations for smooth leg movement. From January 9 to February 9, I wrote and refined the code, implementing a basic walk cycle using inverse kinematics principles.
Throughout the process, I used Fusion360’s measurement and restriction tools to ensure parts fit, and even integrated VEX components to solve missing part issues. I also kept a daily journal and recorded videos to track progress and document my design choices.
Improvement Process
Iteration was key to fixing challenges
3D Printing & Assembly Adjustments：
My first test prints had fitment issues—some parts were too loose or rigid. I reprinted components multiple times, adjusting thickness and joint tolerances. For example, I redesigned the motor bracket to work with all leg segments and added bevels for smoother assembly. I also experimented with a compliant mechanism in the foot to improve stability, though it required several prototypes to get the flexibility right.
Coding & Kinematics Refinement:
Early code produced jerky movement because I used a simple “move-to-position” system. I refined the inverse kinematics logic by sketching joint rotations on paper and testing small code snippets—adjusting angle calculations to reduce stutter. While I couldn’t implement full path planning, I added small delays between position shifts to make the walk cycle smoother.
Structural Durability:
I strengthened the design by interlocking parts and making the base larger to prevent wobbling during movement. These changes came from testing the leg’s weight-bearing capacity—early versions flexed too much, so I increased part thickness in high-stress areas.
Project Achievements
I successfully mastered 3D modeling with Fusion360—now using it in my robotics club to prototype VEX designs—and gained a practical understanding of inverse kinematics, a concept I’d only heard of before. I also improved my self-management and research skills.
-
Mechanical Success:
The final hexapod leg was fully 3D modeled by me and structurally sound—100% of 12 survey respondents said it looked “strong and rugged.”
-
Coding Success:
The code included core inverse kinematics components, meeting my goal of learning and applying the concept. While the walk cycle wasn’t perfectly smooth, 75% of survey respondents rated its fluidity 4-5/5.
-
Aesthetic Success:
The design was well-received—58.3% of respondents gave it a 5/5 for aesthetics, and 41.7% gave 4/5. I prioritized functionality but added parallel support structures and consistent spacing to keep it sleek.
Overall, the project met all my success criteria. It not only built my technical skills but also showed me how walking robots could impact real-world scenarios—like rescue missions or dangerous exploration—reinforcing my interest in engineering.
Science Rollercoaster
