Nvidia

Software Engineer

in a team environment for public-facing site including an event calendar, event timeline, and carousel using Typescript, Astro and Tailwind CSS

achieving a 100/100 lighthouse score, following ADA guidelines and modern SEO practices

Developed and pushed to production 5 web components in a team environment for public-facing site including an event calendar, event timeline, and carousel using Typescript, Astro and Tailwind CSS achieving a 100/100 lighthouse score, following ADA guidelines and modern SEO practices

10-15 personalized lesson plans each week

involving advanced math, Java, and Python

leading to 85% of students earning AP scores of 3 and above

Created and implemented 10-15 personalized lesson plans involving advanced math, Java, and Python each week to facilitate mastery of SAT/ACT math and AP Computer Science, leading to 85% of students earning AP scores of 3 and above

implementing optimized algorithm parameters for improved jet grooming in the Large Hadron Collider

Enhanced the soft drop algorithm by implementing optimized algorithm parameters for improved jet grooming in the Large Hadron Collider, reducing the data processing time of over 5000 files by approximately 10%.

using Python to perform simulations of black holes

Conducted 50+ hours of research on Investigating Black Holes and Imaging Methods using Python to perform simulations of black holes and presented a poster during STEM Day 2024.

MongoDB, and Express to develop a full-stack real estate platform

enabling users to explore detailed information on properties across diverse locations and property types

Harnessed [MongoDB, and Express](https://github.com/kunal7219/Estate-Explorer) to develop a full-stack real estate platform, enabling users to explore detailed information on over 100 properties across diverse locations and property types.

30+ strategic partners, 20% improvement in partner apps’ monthly active users

interactive banner in Google Workspace Marketplace using React/Node-like framework tailored to enhance visibility

yielding a 20% improvement in partner apps’ monthly active users

Engineered interactive banner in Google Workspace Marketplace using React/Node-like framework tailored to enhance visibility for 30+ strategic partners, yielding a 20% improvement in partner apps’ monthly active users.

scalable RESTful API in Java to efficiently query databases for banner apps, optimizing queries

reductions in user wait time and boosted platform reliability

Implemented scalable RESTful API in Java to efficiently query databases for banner apps, optimizing queries and reducing promise response times by 16% which led to reductions in user wait time and boosted platform reliability.

comprehensive testing suites (unit, screenshot, and integration tests) in JUnit and Jest

significantly reducing post-deployment issues

Deployed comprehensive testing suites (unit, screenshot, and integration tests) in JUnit and Jest to rectify critical bugs early in development cycle, achieving 90% code coverage and significantly reducing post-deployment issues.

22% decrease in database lookup and report building times

streamlining data retrieval, caching frequently accessed data, and reducing excess computation

resulting in a 22% decrease in database lookup and report building times

Modernized coroner’s databases in SQL by streamlining data retrieval, caching frequently accessed data, and reducing excess computation, resulting in a 22% decrease in database lookup and report building times.

92% prediction accuracy for toxic messages and 84% prediction accuracy for gender discriminatory messages

Miso via Python Discord Bot and ReactJS site with a RESTful API on Flask

yielded a 92% prediction accuracy for toxic messages and 84% prediction accuracy for gender discriminatory messages

Implemented Miso via Python Discord Bot and ReactJS site with a RESTful API on Flask which yielded a 92% prediction accuracy for toxic messages and 84% prediction accuracy for gender discriminatory messages.

Data Scientist

large datasets, increased data analysis efficiency by 40%

GCP’s BigQuery and SQL to manage and process large datasets, automating ETL Pipelines

increased data analysis efficiency by 40%, providing real-time, accurate data for dynamic dashboards

Leveraged GCP’s BigQuery and SQL to manage and process large datasets, automating ETL Pipelines that increased data analysis efficiency by 40%, providing real-time, accurate data for dynamic dashboards.

advanced Computer Vision (Llava) and NLP (BERT) models

significantly enhancing the accuracy of relevant image retrieval

Integrated advanced Computer Vision (Llava) and NLP (BERT) models, achieving a 25% improvement in Recall@5, significantly enhancing the accuracy of relevant image retrieval.

Implemented under-sampling and ensemble techniques to address class imbalance, leading to 15% improved performance.

Implemented MongoDB aggregation pipelines, reducing query execution times by up to 50%.

enabling users to explore detailed information across diverse locations and property types

enabling real-time scraping with Cheerio and Bright Data

Implemented a search bar for users to input Amazon product links, enabling real-time scraping with Cheerio and Bright Data, processing over 500 products daily.

using Nodemailer to notify users of price drops or stock changes

Email Notifications: Developed an alert system using Nodemailer to notify users of price drops or stock changes, successfully sending over 100 emails monthly, improving user engagement.

$24,000 yearly advertising budget, average 20% decrease cost-per-click

created reports, and queried data from Facebook Pages REST API using Python and employee database using SQL

leading to an average 20% decrease cost-per-click on ad campaigns

Allocated $24,000 yearly advertising budget, created reports, and queried data from Facebook Pages REST API using Python and employee database using SQL leading to an average 20% decrease cost-per-click on ad campaigns

AI tools for research and variable estimation techniques

resulting in data-driven recommendations for stakeholders

Utilized AI tools for research and variable estimation techniques, including market sizing and revenue projection, to predict future industry trends, resulting in data-driven recommendations for stakeholders.