Kusuma Priya Boddu - Software Products & Platform Engineer

Summary

Master's graduate in Data Science with 2+ years of experience in software engineering, analytics, and data-driven solution development. Skilled in designing and implementing end-to-end machine learning, deep learning, and reinforcement learning solutions using Python, SQL, PyTorch, TensorFlow, and Scikit-learn. Hands-on experience developing AI applications in time-series forecasting, computer vision, and predictive analytics. With expertise in feature engineering, model evaluation, reinforcement learning, and model deployment. Passionate about building scalable AI solutions that solve real-world business problems.

Overview

6

years of post-secondary education

3

years of professional experience

Work History

Software Products & Platform Engineer

Accenture Technology

Hyderabad, India

11.2020 - 12.2023

Built automated SQL & Python ETL pipelines, reducing manual reporting efforts by 35% and improving data consistency.
Developed 20+ KPI dashboards in Power BI and Excel enabling leadership to monitor operational performance, trends, and business metrics.
Analyzed 500K+ operational and ERP records using SQL, identifying reporting discrepancies and improving data accuracy by 25%.
Performed exploratory data analysis and anomaly detection, reducing reporting errors by 30%.
Conducted root cause analysis on performance metrics, identified process inefficiencies, and recommended data-driven improvements to enhance operational effectiveness.
Partnered with cross-functional teams (product, operations, and engineering) to define reporting requirements and improve analytics workflows.
Delivered insights and reports to stakeholders, facilitating strategic decision-making across technical and business teams.
Supported data validation and metric tracking, ensuring accuracy of analytics and reporting outputs to uphold data integrity.

Education

Master of Science - Data Science

University of North Texas

Dallas, TX

01.2024 - 12.2025

B.Tech - Electronics & Communication Engineering

JB Institute of Engineering And Technology

Hyderabad

08.2017 - 07.2021

Skills

Programming: Python, SQL

Machine Learning: Scikit-learn, Regression, Classification, Clustering, SVM, Feature Engineering, Model Evaluation, Hyperparameter Tuning, Time Series Forecasting

Deep Learning: PyTorch, TensorFlow, Transformers, ST-GCN, Neural Networks

Reinforcement Learning: Group Relative Policy Optimization (GRPO), Policy Optimization, Sequential Decision Making

Computer Vision: OpenCV, MediaPipe, Landmark Detection, Gesture Recognition

Data Science: Pandas, NumPy, Matplotlib, Data Cleaning, Exploratory Data Analysis (EDA), Statistical Modeling

Deployment & Frameworks: Flask

Visualization: Power BI, Tableau, Excel

Databases: MySQL, Azure SQL, Snowflake

ETL development

PROJECTS

Machine Learning and AI Projects
Demand Forecasting using Group Relative Policy Optimization (GRPO) ASL-Echo: American Sign Language Recognition Obesity Prediction using Machine Learning

Python • PyTorch • Reinforcement Learning • Walmart M5 Dataset

Built a reinforcement learning-based demand forecasting framework using the Walmart M5 retail dataset with over 30,000 products and 1,913 days of historical sales, pricing, and calendar information.
Designed an end-to-end forecasting pipeline including preprocessing, feature engineering, Ridge Regression baseline modeling, and chronological train/test splitting.
Implemented a custom GRPO environment with a 30-day observation window, action-based forecast correction, reward optimization using negative MSE improvement, and replay buffer-based policy learning.
Evaluated forecasting performance using MASE and compared GRPO against Temporal Fusion Transformer (TFT) and hybrid ensemble models, demonstrating competitive accuracy with significantly lower computational cost.

Python • OpenCV • MediaPipe • ST-GCN • Transformer

Developed a deep learning pipeline for American Sign Language recognition using the Microsoft MS-ASL dataset.
Built a preprocessing pipeline with OpenCV and MediaPipe to extract hand, pose, and facial landmarks from videos.
Applied normalization, augmentation, class balancing, and label encoding to improve model performance.
Trained hybrid ST-GCN and Transformer models for gesture classification and evaluated performance using Accuracy, Precision, Recall, and confusion matrix analysis.

Python • Scikit-learn • SVM • Flask

Built a machine learning application to predict obesity levels using health and lifestyle features.
Performed data preprocessing, feature engineering, and trained Support Vector Machine (SVM) models for classification.
Developed a Flask-based web application for real-time prediction and deployed the trained model using serialized artifacts.

Timeline