![](source/img/home-bg.jpg)
Lakshita Shetty
I'm a
About
![](source/img/profile-img.jpg)
Hi there! I'm Lakshita Shetty, a Graduate in Applied Data Science from University of Southern California. I love transforming data into actionable insights and have a strong background in deep learning, data mining, and machine learning. I earned my Bachelor's in Electronics & Telecommunication from D. J. Sanghvi College of Engineering in Mumbai, graduating in the top 5% of my class. I've also completed a Business Analytics and Data Science course in collaboration with IBM.
Currently, I'm a Data Science Intern at Fiscal Inc., where I use Large Language Models to extract financial terms and develop Python code for financial calculations. I've also worked as a Research Assistant at USC Marshall, automating data collection and enhancing data-driven research. Some of my favorite projects include studying gender bias in social media moderation using BERT models, analyzing LinkedIn usage for job success, creating a data-driven F1 racing dashboard, and developing a movie recommendation system.
Outside of work, I volunteer extensively in my community. Whether it's hitting the courts for tennis and squash or diving into a handball match, I balance my love for data with the exhilarating pulse of physical activity and competition.
Thank you for stopping by! I’m excited to continue my journey in data science and see where my curiosity and creativity will take me next.
Skills
Languages:
![](https://upload.wikimedia.org/wikipedia/commons/thumb/c/c3/Python-logo-notext.svg/1869px-Python-logo-notext.svg.png)
PYTHON
![](https://cdn.freebiesupply.com/logos/large/2x/html-5-logo-png-transparent.png)
HTML5
CSS3
![](source/img/skills/JavaScript.png)
JAVASCRIPT
![](source/img/skills/R.png)
R
SCALA
C
![](source/img/skills/C++.png)
C++
![](source/img/skills/java.png)
JAVA
![](source/img/skills/Matlab.png)
MATLAB
Databases:
![](source/img/skills/mysql.png)
MYSQL
![](source/img/skills/sqlite.png)
SQLITE
![](source/img/skills/sqlserver.png)
SQL Server
MONGODB
![](source/img/skills/Firebase.png)
FIREBASE
![](source/img/skills/Oracle.png)
ORACLE
![](source/img/skills/postgresql.png)
POSTGRES SQL
![](source/img/skills/snowflake.png)
SNOWFLAKE
![](source/img/skills/dynamodb.png)
AMAZON DYNAMO DB
Frameworks / Web Technologies:
![](source/img/skills/spark.png)
APACHE SPARK
![](source/img/skills/hadoop.png)
APACHE HADOOP
![](source/img/skills/Bootstrap.png)
BOOTSTRAP
![](http://pluspng.com/img-png/nodejs-logo-png-nice-images-collection-node-js-desktop-wallpapers-370.png)
NODE.JS
![](https://gitlab.com/uploads/-/system/project/avatar/31295677/pngwing.com__1_.png)
DJANGO
![](source/img/skills/pytorch.png)
PYTORCH
![](source/img/skills/keras.png)
KERAS
![](source/img/skills/tensorflow.png)
TENSORFLOW
![](source/img/skills/d3.png)
D3.JS
Cloud/Tools:
![](source/img/skills/docker.png)
DOCKER
![](source/img/skills/aws.png)
AWS
![](source/img/skills/gcp.png)
GCP
![](source/img/skills/github.png)
GIT
![](source/img/skills/Airflow.png)
AIRFLOW
![](source/img/skills/TABLEAU.png)
TABLEAU
![](source/img/skills/powerbi.png)
POWER BI
![](source/img/skills/ibm-cognos.png)
IBM COGNOS
Education
Work Experience
Data Science Intern
Jan 2024 - Present
Fiscali Inc, Los Angeles, CA
- Leveraging the capabilities of Large Language Models to extract and define financial terms from contracts, loan agreements, and financial documents.
- I develop specialized prompts that not only generate these definitions but also facilitate the creation of Python code for accurate financial calculations.
- Evaluated various LLMs to select an optimal open-source model for reverse engineering.
- This process involves refining the model to produce tailored Python code, with an aim to integrate these capabilities into financial software systems effectively.
Research Assistant
May 2023 - Jan 2024
Leventhal School of Accounting - USC Marshall, Los Angeles, CA
- Automated data collection using advanced web scraping techniques, including Python, Beautiful Soup, and Selenium, and used SQL for efficient data storage and querying, resulting in a 50% increase in data acquisition efficiency.
- Utilized SQL to extract and manipulate large datasets from Excel and other sources, enhancing the analysis of financial audit reports located on the EDGAR website and contributing to a 20% acceleration in data-driven research initiatives.
Audio Visual Developer Intern
Jun 2021 - Jul 2021
Pheme Software Pvt Ltd - IBM, Mumbai, MH, India
- Designed and developed E-learning audio and videos based on Technology.
- Storyboarding before the creation of audio and video (interactive, animated, or simple text presentations with voice synchronization)- video editing
- Created content that was used on their educational platform
Finance Intern
Nov 2019 - Dec 2019
ITD Cementation India Ltd, Mumbai, MH, India
- Foundation training in the finance department - learned daily finance tasks and workflow
- Developed familiarity to financial terminology which aided my coursework in Predictive Analysis
- Optimized the system to indicate the due date of LCs and BGs
- Learnt bookkeeping, concepts of cash flow and payroll management
PROJECTS
![](assets/img/woodsy_wayfinder.png)
Gender Bias in Content Moderation and Censorship across Social Media
Conducted a comparative study on gender bias in social media moderation using HuggingFace's Social Bias Frames, to examine demographic influences. Implemented bias mitigation strategies with distilBERT and data augmentation, achieving an impressive 87.57% accuracy and substantial improvements in fairness metrics. • Fine-tuned BERT models to enhance fairness, employing feature swapping and masking to effectively reduce gender biases.
Github Repository
![](assets/img/woodsy_wayfinder.png)
LinkedIn Study - Premium Usage, Networking and Job Success
• Investigated the impact of network activity and premium LinkedIn usage on job search outcomes by surveying over 75 participants. • Conducted independent samples t-tests, correlation tests, and statistical analyses of the research variables. • Evaluated Type II error rate to be 2.4% and Type I error through power analysis using SPSS to determine the proposed solution.
Project Proposal
![](assets/img/csgo.png)
F1 Fanalytics Dashboard
• Enhanced Formula 1 strategic decision-making with a visually compelling user-friendly interface, incorporating KPIs and innovative features like animated line charts, choropleth maps, 3D globes, and Coxcomb charts. • Utilized data analytics and technology stack expertise to create an immersive portal, empowering users to decode data-driven exploration of F1 racing dynamics.
Github Repository
![](assets/img/dicom.png)
Smart Choice - A Movie Recommendation System
• Replicated Hadoop's distributed file system by harnessing Netflix, Amazon, and Disney databases, boosting data management efficiency by 40%. • Integrated partitioned MapReduce with Google Firebase and MySQL, enhancing user engagement by 25% with a customized interactive web app.
Github Repository
![](assets/img/ds.png)
Aura - Emotion Recognition System
Developed a machine learning system employing data augmentation techniques to detect 7 emotive patterns from speech and facial expressions, achieving a 1.5x accuracy enhancement. Created a user-friendly mood-display web app, catering to neuroatypical individuals, notably those with autism, resulting in a 30% increase in user satisfaction • Published a technical paper entitled “Aura - Emotion Recognition System ” in DJ Spark Journal in 2022 with ISBN: 978-93-5593-448-2 and secured 4th position
Github Repository
![](assets/img/database.png)
Off-grid Wi-Fi communication system with Raspberry Pi and Android Application
• Designed & implemented an efficient Off-grid Wi-Fi communication system using Raspberry Pi and Android Application. Raspberry pi was used to create a LAN. Key features and benefits include: Devices could connect to the Wi-Fi, and could chat, share files or even talk to each other; enabling effective communication even in areas that do not support cellular/Wi-Fi networks. • Published a technical paper entitled “Off-grid Wi-Fi communication system with Raspberry Pi and Android Application” in DJ Spark Journal in 2021 with ISBN: 978-93-5437-739-6 and secured 5th position
Github Repository
![](assets/img/dccn.png)
Breast Cancer Prediction Using ML
Designed a breast cancer prediction model which can predict the likelihood of tumor in patients with highest accuracy. To achieve maximum accuracy, they used multiple machine learning classification algorithms such as Decision Tree, SVM and Naïve Bayes, Logistic Regression, K-nearest neighbors, ANN, Kernel SVM, Random Forest Classification, which are all used to detect cancer at a preliminary stage.
Project Proposal
![](assets/img/dccn.png)
Optimisation of Public Transportation - B.E.S.T (Brihanmumbai Electricity Supply and Transport)
Presented a solution to optimize an intricate Public Bus Transportation System. We utilized the knowledge of dashboard visualization, python skills to come up with an advance application-based solution.
Project Proposal
Achievements and Extracurricular Activities
Achievements
Viterbi Student Graduate Mentor Spring 2024
Issued by University of Southern California's Volunteer Center
Viterbi Student Graduate Mentor Fall 2023
Associated with University of Southern California
DISTINGUISHED SERVICE AWARD 2024
Issued by City of Los Angeles and New 9th Council District along with USC’s Volunteer Center
DISTINGUISHED SERVICE AWARD 2023
Issued by City of Los Angeles and New 9th Council District along with USC’s Volunteer Center
Secured 4th position in the state level project based paper presentation competition - DJ Spark 2022
Issued by DJS IETE
Secured 5th position in the state level project based paper presentation competition - DJ Spark 2021
Issued by DJS IETE
Activities
NSS (National Service Scheme) Certificate - 120 hrs of Social Service - 2019-2020
Issued by NSS
EOS 2016 Photo-Journalism - 1st position
Associated with Pace Jr. College
Silver Level Certificate of The International Award for Young People (IAYP)
Issued by The Duke Of Edinburgh's International Award
Bronze Level Certificate of The International Award for Young People (IAYP)
Issued by The Duke Of Edinburgh's International Award
RIO+22 UN Sustainable Energy for All India (2014) - A GRADE CERTIFICATE
Issued by IARC' Centre for United Nations
Taekwondo - RED-I
Issued by Taekwondo Federation of India
Contact
Address
Los Angeles, California, US, 90007
Call Us
+1 213-561-8401
Email Us
lshetty@usc.edu
lakshi276@gmail.com