Siri Yellu
Graduate Research Assistant | MS in Computer Science (Data Science)
Data Science, Machine Learning, Analytics
Turning data into actionable insights through analytics and machine learning.
About Me
Academic Excellence
MS in Computer Science (Data Science) with 4.0 GPA
Machine Learning
Predictive modeling and NLP solutions
Data Analytics
Healthcare analytics and visualization
I am currently pursuing my MS in Computer Science with a concentration in Data Science at Kennesaw State University, maintaining a perfect 4.0 GPA. With a strong foundation in data analytics, machine learning, and software development, I specialize in healthcare analytics, natural language processing, dashboard development, and predictive modeling. My experience includes building ML pipelines for cancer image analysis, developing NLP-powered chatbots, and creating interactive dashboards that transform complex data into actionable insights. I am passionate about leveraging technology and data to solve real-world challenges and am actively seeking opportunities in Software Engineering, Data Analytics, Machine Learning, and Data Science roles.
Projects
A showcase of AI/ML, computer vision, and data analytics projects spanning healthcare, research, and real-world applications
- Built multi-agent RAG system using LangChain and GPT-4 for medical document Q&A
- Integrated vector search with Pinecone for semantic retrieval of research papers
- Deployed interactive Streamlit app with real-time chat interface and code assistance
- Designed Tableau dashboard integrating CDC data with 3K+ U.S. county demographics
- Uncovered healthcare access disparities correlating with 40% higher mortality rates
- Supported public health decision-making with accessible, interactive visualizations
- Built comprehensive dashboard analyzing 50K+ delivery records and fleet metrics
- Identified root causes of 23% on-time delivery gap through geospatial analysis
- Visualized risk zones and bottlenecks improving logistics operations by 18%
- Analyzed 20+ years of NOAA climate data and eBird migration records (500K+ observations)
- Built Random Forest model predicting migration timing with 87% accuracy
- Created interactive visualizations supporting conservation policy recommendations
- Developed Android app with TensorFlow Lite for on-device melanoma detection
- Achieved 92% accuracy using EfficientNet model trained on HAM10000 dataset
- Published research paper and deployed user-friendly mobile interface for clinical screening
- Implemented DeepLabCut-based pose tracking for fruit fly behavioral analysis
- Processed 10K+ video frames with multi-keypoint detection and temporal smoothing
- Enabled quantitative analysis of locomotion patterns for genetics research
- Developed real-time face recognition system with 95% accuracy using dlib and OpenCV
- Implemented liveness detection to prevent spoofing attacks and ensure security
- Built web-based admin dashboard for attendance management and reporting
- Built conversational AI using GPT-3.5 with financial domain fine-tuning
- Integrated expense tracking API and generated personalized savings recommendations
- Deployed Streamlit interface with natural language understanding for financial queries
- Trained BERT-based model on 100K+ code samples to identify security flaws
- Achieved 89% precision in detecting SQL injection and XSS vulnerabilities
- Created VS Code extension for real-time security feedback during development
Publications & Research
Applied AI research with real-world impact across healthcare, NLP, and computer vision
Georgia Tech Global Learning Center
2024 IEEE ICSC Conference, Laguna Hills, CA
E3S Web of Conferences 2024, India
Healthcare AI & Medical Imaging
Investigating Spatial Patterns of Tumor and Stroma in Gastric and Colorectal Cancer for Survival Prediction
Sanghoon Lee, Yellu Siri, Sung Hak Lee, Jae-Ho Cheong, Minji Kim, Sunho Park, Tae Hyun Hwang
IEEE-EMBS BHI 2025
Analyzed spatial tumor-stroma patterns in gastric and colorectal cancer to predict patient survival outcomes using deep learning.
Evaluation of Hand-Crafted Features with Mask Images Obtained from PanNuke Dataset Using Bayesian Optimization and Machine Learning Models
Yellu Siri, S. C. Koganti, H. Elghazzali, Y. Zhao, D. F. Williamson, S. Lee
IEEE ICSC 2025
Optimized hand-crafted features for nuclei classification on PanNuke dataset using Bayesian methods and ML models.
Task-ready PanNuke and NuCLS Datasets: Reorganization, Synthetic Data Generation, and Experimental Evaluation
S. C. Koganti, Yellu Siri, J. Yun, S. Lee
IEEE Access
Reorganized medical imaging datasets and generated synthetic data to improve nuclei segmentation tasks.
Color Normalization Analysis for Semantic Image Segmentation on Histopathology Images
V. Yaganti, S. C. Koganti, Yellu Siri, S. Lee
IEEE SoutheastCon 2025
Evaluated color normalization techniques to improve semantic segmentation accuracy on histopathology images.
Natural Language Processing
Enhancing Sentiment Analysis Accuracy by Optimizing Hyperparameters of SVM and Logistic Regression Models
Yellu Siri, S. Afroz, R. U. Rani
E3S Web of Conferences 2024
Improved sentiment classification performance through systematic hyperparameter tuning of SVM and logistic regression.
Generative AI
Generative Adversarial Quest: Enhancing Question and Answer Generation Through GAN
Yellu Siri, C. V. S. Satyamurty, S. N. Appe
ICICC 2024
Leveraged GANs to automatically generate high-quality question-answer pairs for educational applications.
Computer Vision
Early Stage Identification of Tomato Leaf Diseases using VGG16 and MobileNet CNNs
Yellu Siri, R. Jagarlapudi, S. T. Salehundam, K. J. Mohan, K. V. Sharma
Macaw International Journal of Advanced Research in Computer Science
Developed CNN-based models for automated early detection of tomato plant diseases from leaf images.
Experience
Kennesaw State University
Developing hybrid ML pipelines for cancer image analysis and spatial pattern analytics, combining deep learning with traditional computer vision techniques.
Pathsetter
Built NLP chatbots and improved engagement and response accuracy through advanced natural language processing techniques.
DLRL
Developed computer vision systems and automated dashboards for image processing and analysis workflows.
Skills & Technologies
Get In Touch
I'm always open to discussing new projects, creative ideas, or opportunities to be part of your vision.
Let's Connect
Feel free to reach out through any of the following channels. I typically respond within 24 hours.