Ravi Teja Challa

As a former AWS Engineer at Wipro, I've cultivated a robust skill set in AWS services, spanning EC2, S3, Lambda, VPC, and RDS. My focus on scalability, security, and performance led to the successful implementation of cost-effective solutions. Complementing this hands-on experience, I dedicate time daily to enhance my problem-solving skills in Java through LeetCode. This dedication, coupled with my academic pursuit of an MSCS at UAB, positions me to thrive in the dynamic realm of computer science.

AWS Certified Solution Architect - Dec 2022-Dec 2025

AWS Certified Cloud Practitioner - Mar 2022-Mar 2025

JSP, Servlets and JDBC, Udemy - Sep 2022

Full Stack Java - Aug 2021

Web Development Bootcamp - Oct 2020

Telecom Customer Attrition Analysis - Aug 2023 — Dec 2023

• Led end-to-end telecom attrition project, utilizing advanced analytics and ML. Processed 7043 customer records from Kaggle, conducting EDA, cleaning, and outlier analysis. Employed feature extraction, encoding, and algorithms (Logistic Regression, Adaboost, Decision Tree) for an 81% model accuracy. Evaluated performance with ROC curves, highlighting variable impact. Applied ensemble techniques (Adaboost, Bagging, Voting) and 5-fold cross-validation, achieving a 49% recall rate for churning detection.

Deep dual domain joint discriminant feature framework for  emotion-based music player -Springer (https://link.springer.com/article/10.1007/s13198-024-02382-z)

      • Proposed a deep dual-domain joint feature framework for facial emotion recognition, integrating linear discriminant analysis. Leveraged pre-trained models (EfficientNet, Inceptionv3, ResNet) for face detection and emotion pattern learning. Published successfully in Springer Journal on 29th May 2024 

Facial Emotion Recognition Using Custom CNN Architecture and Pretrained Models - may 2024

      • Developed a CNN-based model to recognize seven facial emotions: anger, disgust, neutral, fear, happy, sad, and surprise. Trained and validated on the FERC2013 dataset. Compared model performance with pre-trained models like ResNet50 and VGG16, achieving improved accuracy through architectural modifications. Enhanced emotion-sensitive HCI systems for applications in surveillance, commerce, and education

Rank predictor  - Jan 2022

     • Created a website using Bootstrap, CSS, and JavaScript with a predictive model for university admission based on academic performance. Implemented diverse machine learning algorithms, achieving tested accuracy