Education
California State University, Los Angeles, USA
Master of Science in Computer Science
Jan 2023 - Dec 2024
GPA: 3.93/4.0
Publication
A Hybrid Pub/Sub and REST-based Communication Strategy for Intelligent IoT System Interoperability
Shivani M Kotian, Manveen Kaur, Youssef Elzein, Andres Dominguez, Jung Soo Lim, Navid Amini
IEEE CCNC 2025 - IoT Interoperability and the Web of Things (IIWOT) Workshop
View PublicationPES University, Bangalore, India
Bachelor of Engineering in Computer Science
Aug 2016 - Sept 2020
GPA: 3.64/4.0
Publication
Epigraphiology: A Hybrid Approach for Measuring and Analyzing Influence Diffusion in Article Networks
Sudeepa Roy Dey, Shivani Kotian, Anmol Agarwal, Arshika Lalan, Gambhire Swati Sampatrao, Snehanshu Saha
Journal of Scientometric Research
View PublicationProfessional Experience
Graduate Research Assistant - California State University, Los Angeles, USA
Jun 2023 - Present
- Developed a hybrid communication strategy for IoT systems using MQTT and REST based pub/sub protocols
- Conducted performance analysis of object detection CNN models on NVIDIA Jetson-Orin, Intel NUC, and Raspberry Pi, processing datasets of 10,000+ images to evaluate computational efficiency for real-time IoT systems
- Assessed training time, inference time, and hardware utilization metrics to validate suitability for UAV-assisted ITS applications.
- Published findings in IEEE CCNC 2025 - IIWoT Workshop, addressing QoS challenges in resource-constrained environments.
Data Scientist - Ernst & Young, Bangalore, India
Oct 2020 - Jan 2023
- Engineered Advanced Forecasting Models: Built ARIMA, Croston, RandomForest, and XGBoost models with backtesting, increasing FMCG sales volume prediction accuracy by 15% over traditional methods.
- Developed Automated ETL Pipelines: Utilized SQL and PySpark to handle 500GB+ of data, reducing extraction time by 60% and supporting 10+ weekly forecast runs across multiple markets.
- Optimized End-to-End Model Lifecycle: Leveraged Microsoft Azure to cut processing time by 70%, contributing to €4.2 million in cost savings through efficient demand planning and cross-functional feature engineering.
- Streamlined Analytics & Reporting: Automated KPI dashboards in Power BI, performed ad-hoc data requests, and produced Root Cause Analysis reports—reducing manual efforts by 70%, accelerating KPI monitoring by 25%, and earning EY Kudos! awards in 2021 and 2022 for exemplary performance.
Intern - Ernst & Young, Bangalore, India
Jan 2020 - Feb 2020
- Applied advanced deep-learning and computer vision methods to identify barcodes on supply chain pallets, significantly enhancing warehouse tracking and operational efficiency.
- Augmented the dataset with transformations to increase the dataset size to 3k+ images and utilized Faster R-CNN with ResNet50 for detection, followed by pyzbar for decoding, boosting model accuracy.
- Achieved a 95% success rate on diverse barcodes, streamlining supply chain operations and improving inventory tracking efficiency.
Data Science Projects & Contributions
CodeSage
Multi-Modal Question-Answering System
This project is a RAG-based Q&A system that answers software engineering questions using textbook knowledge, featuring multi-modal document analysis (text/images), semantic caching with FAISS/Chroma vector stores, and a Streamlit interface powered by Groq's Llama-3 model.
CerebroVision
Advanced CNN Diagnostics
Employed transfer learning with ResNet50, VGG19 for brain tumor classification, achieving 99.88% accuracy.
Decoding Dreams
Sleep Efficiency Analysis
Analyzed sleep efficiency datasets to explore affecting factors using data visualization techniques.