Sakshi Roongta

I am currently a software engineer at Amazon. I studied at Carnegie Mellon University, majoring in Computer Science and Business Administration. I am always interested in learning about new opportunities to work on teams to solve complex problems that make a positive impact. If you are interested, you can contact me at the email above. I look forward to connecting with you!


Experience

Software Development Engineer

Amazon

  • Worked across teams to develop a new set of APIs by having discussions to collect requirements from external teams, designing models for the APIs, and getting them approved by multiple stakeholders.
  • Investigated design tradeoffs for these APIs by analyzing different compute options such as Lambda and ECS by looking at factors such as cost, auto-scaling, instance management, and invocation impact on latency, investigating multiple storage options for our audit information including S3, CloudWatch, and OpenSearch, investigating whether batch processing for the APIs would be an issue based on forecasted payload sizes, and designing how authorization, auditing, idempotency, etc. would work.
  • Led the development of these major new features by creating an implementation plan and tasks in each sprint. Wrote code in Java using AWS CDK, Swagger, and OpenAPI and deployed them using a CI/CD pipeline. Documented guide on how to use APIs and received positive feedback from clients.
  • Designed and helped implement a new user interface for Amazon’s fulfillment orchestration system using CloudFront, S3, Java, TypeScript, and React. Project reduced on-call load by around 30%, enabling the team to take on more clients and scale faster.
  • Analyzed ways to reduce our latency spikes in certain APIs and considered various ambiguous potential solutions such as caching, request timeout settings, setting the APIs to be eventually consistent, etc. and implemented solution that successfully eliminated the spikes.
  • Re-designed metrics and alarms to reduce the number of false tickets from 2-3 a week to nearly 0. Ran load tests and analyzed results to understand our max TPS and ensure we were sufficiently scaled to handle peak traffic that occurred during the holidays.
  • Reviewed code developed by other developers and provided feedback to ensure best practices. Addressed gaps in others code reviews related to functionality, testing, efficiency, extensible code, and clear variable and file naming conventions.
  • Helped fix high severity issues during on-call by first performing actions to immediately unblock clients and then creating and implementing long-term fixes to prevent the same issues from happening again.

July 2021 - Present

Software Development Engineer Intern

Amazon

  • Designed and implemented an application with a partner that improved user experience when onboarding internal clients to the Amazon fulfillment system using React, Spring, and ECS.
  • Created an AWS Lambda application in Java to authenticate users.
  • Wrote unit and integration tests that provided 90% code coverage using Junit and Mockito.

June 2020 - Aug. 2020

Research Assistant

Carnegie Mellon University, Institute For Software Research

  • Designed and developed an application in C++ to simulate a four-way intersection for autonomous vehicles using vehicle-to-vehicle communication.
  • Induced and studied delays and failures in network and how to maintain safety despite those delays.

Github Link
June 2019 - Aug. 2019

Software Engineer Intern

Infosys

  • Built a proof-of-concept web application using Python, Django, and SQL to help clients create and manage a stock portfolio.
  • Researched financial papers to develop algorithm that calculates a stock’s momentum and z-score to create a well-balanced portfolio.

Github Link
June 2018 - Aug. 2018

Tartan Student Fund

Vice President

  • Managed fifty analysts to prepare equity reports and pitches by completing valuations and analyzing company financial data.
  • Helped manage a portfolio of $50,000 of endowment money for the university.
  • Collaboratively pitched long-term equity investments in the technology, media, and telecommunication sectors.

Aug. 2018 - Aug. 2019

Projects

Social Scrubber: Tartan Hacks Winner

Social Scrubber is a Chrome extension that identifies which parts of a Facebook post may contain personally identifiable information (PII).

Our application detects when the user is making a Facebook post. Users install our application as a Chrome Extension which detects when the user is making a Facebook post. It automatically identifies potential PII in the process of making the post, both in the text box as well as within the uploaded photos. When it identifies potential PII, it warns the user via a pop-up containing the offending information.

ASL Robotic Hand: Build18 Hackathon Winner

The ASL Robotic Hand is a two-way communication system between English and American Sign Language (ASL).

A robotic hand performs the ASL alphabet by taking in input text from a computer. The robot hand uses torsion springs that allows the fingers of the hand to be straight. The strings are connected to the base of the string using strings connected to servo motors. These motors pull the strings and thus the fingers of the hands down and to the left and the right, allowing for movement of the hand.

A Leap Motion sensor takes in hand position and movement to translate ASL to English. We then transformed the data to calculate consecutive fingertip distances, finger bend angle, and knuckle slope. We used Linear Discriminant Analysis to recognize the sign the user was creating.

TL;DR: Sandia National Labs' Choice at Tartan Hacks

TL;DR is a Chrome extension that compiles and summarizes the most notable sections of End User License Agreements (EULAs).

It finds the most notable sections of such a legal document — namely, sections which seem unusual and do not align with typical EULAs — and compiles them into a brief summary. Since it is a Chrome Extension, users can use it directly from their browser when they encounter one of these legal documents. A user would highlight the text they wish to be summarized and would subsequently see a brief summary that lists the most noteworthy parts of the document.

Chord Buddy

Chord Buddy is an application to help beginner guitar players learn and memorize chords on the guitar. It cycles through chords that the user must play within a few seconds and then tells them if the shape and sound of the chord they played is correct using neural networks to recognize the chord's sound and shape on the fingerboard.

A picture and sound recording is taken of the user playing the chord. The picture is run through a VGG neural network model with four convolutional layers and two dense layers. We added a flatten and dropout layer to prevent overfitting since we have limited data and we used SoftMax activation and categorical cross entropy loss. We trained the model on our data using an augmentation configuration to prevent overfitting.

Moodify

Moodify is a terminal application that uses machine learning to create personalized Spotify playlists with up to 30 songs that match a certain mood.

It uses machine learning to classify the five hundred most recent tracks saved to a user’s profile into four possible moods: happy, sad, angry, and relaxed. The mood of the user is obtained using either by taking a picture of the user’s face and reading their emotions or by them manually inputting a mood.

Sailaway: Build18 Hackathon

Sailaway is a VR boating game with a moving platform that responds to the movement of the VR controllers.

It is a project created for Build18, a weeklong hardware hackathon at Carnegie Mellon University. The objective of the game is collect the coins floating around in the water. The person sits on a moving platform we created that is similar to a turntable. When the person turns the boat, the platform will change direction and speed based on the movement of the VR controllers. This allows the user to feel like they are really sitting in a boat and riding on the waves.


Coursework

AI
  • Deep Learning
  • Machine Learning
  • Artificial Intelligence
Algorithms
  • Algorithm Design and Analysis
  • Parallel and Sequential Algorithms
  • Theoretical Computer Science
  • Complexity Theory
Systems
  • Distributed Systems
  • Computer Systems (x86)
Business
  • Negotiation and Conflict Resolution
  • Business Presentations
  • Business Communications

Interests

In my free time, I like to do silly projects with friends. We’ve programmed our Anki Vector robot to be sassy, built an unnecessarily complicated Secret Santa gift exchange, and built a program to settle once and for all whether or not there is a Set left among the twelve cards we’ve laid out when playing the card game Set.

I also love to draw and find it really calming to get lost in and focus on a drawing for a few hours.