Portfolio - Sathvika Reddy Reddem

Sathvika Reddy Reddem

Software Developer

Resume

1

Python Project

About Me

I’m Sathvika, a passionate Computer Science graduate with a deep interest in building impactful software and AI solutions. From developing real-time ultrasound nerve segmentation using U-Net to analyzing global CO2 emissions with LSTM networks, I love solving real-world problems through intelligent systems. I’ve completed industry-recognized courses like CS50, Machine Learning by Andrew Ng, and Python for Everybody, and I enjoy blending creativity with logic. With strengths in leadership, public speaking, and product thinking, I aspire to build solutions that not only work but make a difference. Currently, I’m seeking opportunities to work on innovative tech or AI projects that challenge me, help me grow, and contribute meaningfully to the world.

Key Skills

Python, Java, C, MYSQL, DBMS, LLMs, NLP, MS Excel, HTML5, CSS3, JavaScript, Bootstrap

My Projects

Real-Time Ultrasound Nerve Segmentation using U-Net

Domain/Function: Deep Learning

My Experience

Bharat Intern

Machine Learning Intern (Sep 2023)

Designed a regression model to predict wine quality using scikit-learn.
Implemented preprocessing, feature engineering, training, and evaluation pipelines.
Strengthened understanding of model tuning and supervised learning.

Awards & Certificate

Excel: Mother of Business Intelligence

Get Job Ready: Power BI Data Analytics for All Levels 3.0

English Empower Level B2 – Cambridge

Data Analytics Job Simulation – Forage

Data Analytics with Python – NPTEL

Let's Connect

Feel free to get in touch with me. I am always open to discussing new projects, creative ideas or opportunities to be part of your visions.

Download Resume

Resume

Mail Me

[email protected]

https://www.linkedin.com/in/reddem-sathvika-reddy/

Call Me

+91 8143505868

GitHub

https://github.com/RSathvikareddy

Python Project

Real-Time Ultrasound Nerve Segmentation using U-Net

Domain/Function: Deep Learning

Built a U-Net model to perform real-time segmentation of nerves in ultrasound images using a Kaggle dataset with 5000+ images. Achieved 98.98% accuracy, Dice coefficient of 0.4194, and IoU of 0.2653 using TensorFlow, Keras, and scikit-learn.

View Project
on GitHUB

Other Tools

Python

TensorFlow

Keras

OpenCV

Training logs showing high accuracy (up to 98.98%) and low loss over 30 epochs, along with sample input, ground truth, and segmentation prediction.

Graphs of model accuracy and loss trends with sample predictions showing segmentation performance, including accuracy and Dice metrics.

Project Details

Developed a deep learning-based solution to segment nerves in ultrasound images using a U-Net architecture. The project aimed to assist in real-time medical diagnostics by accurately identifying nerve structures, which are often difficult to detect due to noise and low contrast in ultrasound imaging. Utilized a Kaggle dataset containing 5000+ labeled ultrasound images. The model was built using TensorFlow and Keras, with additional preprocessing and evaluation support from scikit-learn. Achieved a test accuracy of 98.98%, a Dice coefficient of 0.4194, and an Intersection over Union (IoU) score of 0.2653. The project demonstrates the practical application of semantic segmentation in the medical field and highlights the potential of AI in enhancing diagnostic precision. This work also involved hyperparameter tuning, data augmentation, and model evaluation to ensure robustness and generalizability in real-world scenarios.