Hi! I'm Tanvi.


I am a software engineer working at Salesforce under the Data Cloud team passionate about creating in any way I can.

I studied at UC Berkeley where I majored in Computer Science with a focus on data science and how we can use it in a million different ways. I’ve done a lot of cool projects and research in a wide variety of specializations, and have experience working with LLMs/NLP with the Berkeley Speech Group, Machine Learning & Data Science with friends at Big Data @ Berkeley, CI/CD @ Salesforce, and full-stack applications at various different grassroots startups.

I have prior experience working with languages like Python, Java, C, Javascript, Groovy, and several others.

When I'm not doing this, you can find me painting, playing basketball, or collecting cool trinkets˚⋆𐙚｡⋆𖦹.✧˚

Your Profile Picture

My Work

SWE Intern - Salesforce

May 2023 - August 2023

Collected, stored, provisioned unstructured data via the ELK stack from an internal tooling service using tools like Jenkins, Prometheus, and Grafana. Stored data into an AWS Time Series Database, cleaned 100,000+ datapoints and made the data available to pipeline engineering developers to ensure their services are functioning as expected.

SWE Intern - Salesforce

May 2022 - August 2022

Built several APIs to proxy requests from authenticated local users to AWS resources like Secrets Manager and S3, authenticated Salesforce developers using GitHub access tokens. AWS resources were provisioned for multiple master accounts per user group, and User group management and user preferences were stored in MongoDB.

Undergraduate Researcher - Berkeley Artificial Intelligence Research [BAIR]

October 2023 - Present

LLMs and Speech @ Berkeley Speech Group under Gopala Anumanchipalli and Akshat Gupta- Researching a novel approach in knowledge editing LLMs; verifying the persistence of an added fact and related knowledge to an LLM through past endeavors such as ROME, MEND, and MEMIT.

Project Manager - Big Data at Berkeley

August 2021 - Present

Valuenex

Designed a web scraping tool using Selenium to gather recent tweets and Google results about startups and used the chatGPT API to predict evaluation using that info. Built a model around the normal distribution to predict the uniqueness of the company to add to Valuenex’s product “startup finder”

American Eagle

Built an optimization model to find the most efficient assortment of products for a clothing store distribution center to minimize shipping costs. To achieve this, our team created vector embeddings from SKU descriptions and optimized them via a linear program.

Intuitive Surgical

Trained a Computer Vision model to filter blurry images from a medical data set and used PCA/SVM on transformed images via the Laplacian process. This cleaned dataset was used to run an unsupervised model via clustering/PCA to find patterns in the medical data.

Software Developer - Flyerstack

August 2021 - May 2022

Led the development of a full-stack, cross-platform app via React Native for a student-led startup that enabled students in Berkeley to receive notifications about events in and around the campus; used Firebase as the backend for all users & event-related data.

WWDC Scholar - Apple

May 2019

Created a playground using Swift and participated in several interviews (Bustle, Refinery29, iJustine) with Tim Cook; 1 of 350 global participants.

Skills

Python

C

JavaScript

Java

HTML5

CSS3

React Native

Numpy

Pandas

Pytorch

MongoDB

Docker

Golang

SQL

SQLite

NoSQL

Node

AWS

Jenkins

Grafana

Fun Projects!

© Tanvi Khot, 2024