Hi, I'm Vivek!
I'm an undergrad at UC Berkeley studying Computer Science and Applied Mathematics. I have interests rooted in anything problem solving - from tackling research questions, to my content creation adventures.
My reserach interests broadly fall under using large language models to uncover lingustic patterns within text. I currently work at BAIR under the Berkeley NLP Group. My latest work includes developing a state-of-the-art model for detecting AI-generated text. Some of my other work consists of researching entropy trends within documents, which you can find here.
Previously, I've worked at Waymo, where I developed a method for detecting "near" collision events when simulating self-driving cars, and Google, where I developed a weighted linear regression algorithm for clock synchronization.
I also have a soft spot for content creation, creating 15+ animated math videos over at https://youtube.com/vcubingx visualizing topics such as neural networks, complex analysis and fractional calculus. My videos have gathered 75,000+ subscribers and 2,000,000+ views. I've also worked with 3blue1brown in the past, creating interactive lessons on 3blue1brown.com.
Outside of work, I love solving Rubik's cubes and playing badminton. Feel free to contact me at vivekverma [at] berkeley [dot] edu if you want to chat about anything!