About Me
About Me
I'm Palak Kakani👋
Passionate data driven problem solver with over two year of experience building scalable data pipelines, deploying machine learning models, and creating interactive dashboards across healthcare, education, fintech, and AI research. I am pursuing my Masters in Data Science and Analytics at Georgia State University, supported by a strong foundation in Information Technology and a deep interest in real world applications of AI.
I have gained hands on experience with organizations including Mercedes Benz USA, Shepherd Center, Better Business Bureau, Omdena, and IIT Patna. My work spans intelligent document processing, RAG systems, machine learning, generative AI, computer vision, and time series forecasting, enabling data driven insights and measurable impact at scale.
I also enjoy building technology that supports communities. I earned third place at HackHers for StrongHer, an AI wellness companion for single mothers, and created Talent X AI, a voice enabled multi agent career exploration platform during the AI for Good Hackathon hosted by the Institute for Insight with Travelers and Ushers New Look.
Core Technical Skills:
- Languages & Tools: Python, R, SQL, Power BI, Tableau, Git, Docker
- Libraries & Frameworks: Scikit-learn, TensorFlow, PyTorch, Streamlit, Flask
- Cloud Platforms: AWS, Azure, GCP
- Concepts: NLP, Time Series Forecasting, GenAI, CI/CD, ETL, API Integration
- Email:palak11.kakani@gmail.com
- Location:Atlanta, Georgia, USA
Projects
Projects
Python
Streamlit
AWS Bedrock
DynamoDB
Amazon S3
Voice AI
Talent X AI – Voice-Enabled Career Exploration Platform
Python
Streamlit
LangChain
RAG
FAISS
GPT-4o-mini
StrongHer – AI Wellness Companion (3rd Place Winner)
Python
SQL
Visualization GCP (BigQuery, Looker Studio)
Data Analysis, Alpha Vantage API
Stock Market Analytics
Python
Flask
Streamlit
OpenAI GPT-4
Firebase
RAG
EduMentor AI-Powered Learning Assistant
Python
Deep Learning
Visualization
Data Analysis
Medical Image Segmentation
Python
Machine Learning
Visualization
Data Analysis
Stock Market Dashboard
Python
Machine Learning
Visualization
Data Analysis
Identifying Diseases in Chest X-Rays
The project aimed to address limited chest disease diagnostics and healthcare professional shortages in Myanmar by developing an AI-powered solution for early detection. A literature review refined goals, and a diverse dataset, including COVID-19 cases, was collected. Robust deep learning models achieved high accuracy in identifying chest diseases, with a focus on COVID-19 cases.
Python
Machine Learning
Data Analysis
Predicting Student Success
Extensive Exploratory Data Analysis, coupled with feature engineering using Python tools, contributed to the development of a predictive model. This model utilized machine learning algorithms like logistic regression and decision trees. Derived from meticulous data analysis, the model successfully captured the intricate factors influencing student performance.
Python
Computer Vision
NLP
Data Analysis
Image Captioning with Transformer-based Model
Developed an advanced image captioning model, seamlessly blending computer vision and natural language processing. Meticulously preprocessed the MS-COCO dataset, tokenized captions, and integrated a Vision Transformer with a Transformer model. Through iterative fine-tuning and the implementation of custom loss functions, crafted a highly sophisticated model achieving remarkable accuracy. The outcome is an image captioning solution capable of generating coherent and contextually relevant captions for diverse images.
Python
Power BI
Tableau
Data Analysis
Dashboard for Sales of Advertisement Campaigns
Utilized Microsoft Power BI and Excel for comprehensive data analysis, revealing underperforming Campaigns due to low reach, high CPC, and inadequate engagement. Recommendations supported by a detailed Power BI dashboard led to strategic decision-making, enhancing GlobalShala's marketing cost-effectiveness.
Python
Data Analysis
Machine Learning
Weather Forecast
This weather prediction project uses historical data to forecast tomorrow's temperature. It involves downloading and customizing a local weather dataset, cleaning the data, and building a robust model. The systematic approach includes implementing a testing framework for model evaluation and enhancing accuracy by adding predictors. The result is a comprehensive system for precise next-day temperature predictions, showcasing proficiency in data analysis and machine learning.Skills
Skills
Machine Learning & AI
Machine learning, Deep learning, computer vision, natural language processing, retrieval-augmented generation, agentic AI systems, multi-agent workflows, generative AI.
Programming Languages
Python, SQL, R, C/C++, PHP, and API development for data and AI systems.
Data Engineering
ETL pipelines, data modeling, big data processing, Azure Data Factory, Apache Spark, Hadoop, workflow automation, and CI/CD practices.
Databases
MySQL, PostgreSQL, SQL Server, MongoDB, Snowflake, and DynamoDB.
Cloud Platforms
AWS, Azure, Google Cloud, Snowflake, and Databricks for scalable data and AI solutions.
Visualization & Reporting
Power BI, Tableau, Looker Studio, Excel dashboards, and data storytelling for insights and executive reporting.
Tools & Development
Git, GitHub, Docker, VS Code, Jupyter, Agile, Scrum, JIRA, and Confluence.
Background
Background
My Education
Georgia State University
J Mack Robinson College of Business
Aug 2024 - Dec 2025
Masters in Data Science & Analytics
- GPA : 4.21/4.00
- Courses: Data Programming, Data Management, Statistical Foundations, Communicating with Data, Scalable Data Analytics, Predictive Analytics, Machine Learning for Analytics, Deep Learning & Gen AI
Shri Govindram Seksaria Institute of
Technology & Science
Oct 2020 - Jun 2024
BTech in Information Technology
- GPA : 8.86/10.00
- Courses: Data Structures & Algorithms (C++), Data Mining and Warehousing, Database Management Systems, Object-Oriented Programming, Mathematics, Cloud Computing, Operating Systems, Software Engineering, Computer Networks.
Chameli Devi Public School, India
Jul 2018 - Jun 2019
12th Standard
- Percentage: 90%
- Courses: Physics, Chemistry, Mathematics
Chameli Devi Public School, India
Jul 2016 - Jun 2017
10th Standard
- CGPA: 10.00
My Experience
Mercedes-Benz, Atlanta, GA
Aug 2025 – Present
Research Assistant
Data Scientist
Data Scientist
- Built an end-to-end RAG pipeline using LangChain, FAISS, OpenAI Embeddings, and RecursiveCharacterTextSplitter to ingest 58+ dealer reports (PDF, TXT, CSV), generating 176 vectorized chunks and reducing manual review time by 70%.
- Delivered an interactive Streamlit-based UI presenting structured, explainable action plans with 90%+ summarization accuracy, including source citations, confidence scores, and report metadata (effort level, owner role).
- Enabled scalable insights across 100+ Mercedes-Benz dealers, standardizing reporting workflows and improving decision-making time for regional managers.
Shepherd Center, Atlanta, GA
Jan 2025 – May 2025
Research Assistant
Data Scientist
Data Scientist
- Engineered time-series classification models on 40,000+ multimodal MS patient records (demographics, MRI, treatments), achieving 59% G-mean and 77% specificity in predicting physical outcomes (improvement, stable, worsening), supporting early tailored treatment.
- Built scalable ML pipelines using Python, scikit-learn, and XGBoost to process 35,000+ longitudinal treatment records, incorporating 6-month resampling and outlier handling for robust data preprocessing.
- Implemented ARIMA forecasting and SHAP explainability techniques, boosting model accuracy by 22% and enhancing rehab success prediction by 15%.
Better Business Bureau, Atlanta, GA
Aug 2024 – December 2024
Research Assistant
Data Scientist
Data Scientist
- Developed and fine-tuned generative AI models on 3.4M+ Better Business Bureau complaints, leveraging LLMs and Llama-based models via Hugging Face, achieving 81% cosine similarity and outperforming baseline models by 31% (previously 50%).
- Improved AI-generated customer responses, boosting alignment with expected outcomes from 36% to 77%, significantly enhancing response consistency and customer satisfaction.
- Integrated ChatGPT and advanced LLMs into customer service workflows, reducing response time by 40% and streamlining feedback processing, achieving 84% similarity for 5-star ratings and 83% for positive reviews, effectively decreasing human workload while maintaining high satisfaction.
Omdena
Jul 2023 – Sep 2023
Machine Learning Collaborater
Volunteer Project
- Led a multidisciplinary team of 80 contributors in developing a Deep Learning model for chest X-ray disease detection (including COVID-19), achieving an impressive accuracy of ~88.47%.
- Spearheaded data collection, EDA, preprocessing,model improvement, and user interface development.
- Gained hands-on experience in medical image analysis, deep learning, and large dataset handling.
Omdena
Jul 2023 – Aug 2023
Machine Learning Collaborater
Volunteer Project
- Collaborated with 50 team members to develop a data-driven model for predicting student success in the Turkish education system.
- Analyzed diverse student data including demographics, socio-economic status, learning environments, and performance metrics.
- Implemented and compared various ML algorithms (linear regression, decision trees, SVM, ANNs) achieving a 91.29% accuracy in predicting student success.
- Sharpened skills in machine learning, data visualization, EDA, and data analysis through active project participation.
Excelerate (Powered by
Saint Louis University)
Jul 2023- Aug 2023
Data Visualisation Intern
- Led data analysis & visualization team at Saint Louis University, driving 29% cost savings through efficient resource optimization.
- Mastered Tableau to unlock strategic insights and deliver impactful presentations to stakeholders for 13 diverse ad campaigns.
- Fostered a collaborative environment, ensuring successful project outcomes and earned a $500 Excelerate Scholarship.
Excelerate (Powered by
Illinois Institute of Technology)
Jun 2023 – Jul 2023
Data Analyst Intern
- Spearheaded a 20-member team at Illinois Tech Excelerate program to analyze GlobalShala's marketing data for 11 diverse ad campaigns, driving an 80% improvement in campaign performance (measurable metrics: conversion rate, click-through rate) contributing to a $30,000-$40,000 cost saving through optimized campaign strategies.
- Crafted captivating Power BI reports and actionable insights, optimizing future campaigns.
- Mastered data analysis and presentation skills, leveraging Power BI to deliver impactful findings, earning a prestigious $1000 scholarship and badge.
Indian Institute of Technology Patna
Jun 2023 – Jul 2023
Research Intern
- Developed automated catalog generation system for efficient object identification & cataloging in densely packed images (e-commerce, inventory management).
- Leveraged RoboFlow for meticulous annotation of individual products within 700-800 images featuring complex, crowded product arrangements.
- Applied proficient PyTorch skills to construct a YOLOv7 object detection model, resulting in an accuracy of 87.93%, showcasing its effectiveness in effectively identifying objects even in densely packed scenes.
- Gained practical skills in training and refining deep learning models, resulting in accurate detection of catalogs in images
Muskurahat Foundation
Feb 2023 - Mar 2023
Summer Intern
- Actively worked to enhance the education and well-being of underprivileged children while successfully raising Rs 1111 in support of this noble cause.
- Developed fundraising, marketing, and communication skills while creating impactful change.
Certificates
Certificates
Achievements
Achievements
StrongHer – AI Wellness Companion for Single Mothers
🏆 3rd Place Winner – HackHers 2025 (Girls Who Code, GSU)
Oct 2025 • Georgia State University
Built during a 24-hour all-women hackathon, StrongHer is an AI-powered wellness companion designed to support single mothers across the U.S. with parenting guidance, mental health tools, nutrition planning, and access to local resources.
- Developed a RAG chatbot using LangChain, GPT-4o-mini & FAISS for real-time, trusted responses.
- Designed mood tracking, journaling, meal planning, and mindfulness tools in a mobile-first Streamlit app.
- Curated static datasets for childcare, food banks, safety resources & financial help — ensuring privacy and offline access.
- Focused on simplicity, empathy & user-centric flow for underserved caregivers.
Talent X AI – Voice-Enabled Career Exploration Platform
🎯 Built at a 24-Hour Hackathon - AI for Good
Nov 2025 • Hosted by GSU’s J. Mack Robinson College of Business, with partners Travelers and Usher’s New Look.
Talent X AI is an immersive platform enabling users to explore career paths through interactive simulations, multi-agent conversations, 3D avatars, and a voice-enabled AI assistant.
- Integrated AWS Bedrock multi-agent architecture (Master Agent + Profile, Skills, Pathway, Portfolio Sub-Agents).
- Built backend workflow for conversation storage using DynamoDB & portfolio asset management in AWS S3.
- Designed clean data structures enabling real-time agent routing & optimized reasoning flow.
- Enabled voice input, audio playback & avatar-based career exploration using Streamlit UI.
Contact
Get in Touch
Connect with Me
Email: palak11.kakani@gmail.com