Arjun.R.Pulugu

Hi, I'm Arjun Reddy Pulugu.

A

Turning Data into Decisions: Boston's Data Science Innovator, Crafting Insights that Drive Success!

About

I am a Data Analytics & AI graduate student at Northeastern University in Boston. I enjoy extracting insights from complex datasets and developing innovative solutions. Always strive to bring 100% to the work I do. I have worked on technologies like Python, SQL, R, PySpark, Tableau, PowerBI, and various machine learning libraries during my studies and professional experience. I have 19 months of work experience as a Systems Data Analyst at Infosys Ltd, which helped me strengthen my skills in data analysis, visualization, and infrastructure optimization. I am passionate about leveraging AI and machine learning techniques to solve real-world problems impacting millions of users and drive business growth.

Languages: Python, SQL, R
Databases: MySQL, PostgreSQL, MongoDB, ChromaDB, FAISS
Libraries: Tidyverse, NumPy, SciPy, Pandas, Matplotlib, Scikit-learn, OpenCV, NLTK
Frameworks: PyTorch, Keras, TensorFlow, LangChain, ETL
Tools & Technologies: MS Office Suite, Tableau, PowerBI, Git, Docker, AWS, Azure, RAG, Snowflake
Certifications: AWS Certified Cloud Practitioner, AWS Certified Solutions Architect - Associate

Seeking a challenging data-focused position that leverages my skills in Data Analysis and Machine Learning. I am eager to contribute to projects that offer professional development, stimulating experiences, and opportunities for personal growth in the field of data science and analytics.

Experience

Northeastern University

Graduate Teaching Assistant

Provided in-depth guidance on R programming, helping students develop proficiency in data manipulation, statistical analysis, and visualization techniques using real-world datasets.
Facilitated weekly discussions and tutoring sessions on fundamental statistical concepts, including probability, measures of central tendency, and variance, ensuring students could apply these principles in their data analysis projects.
Offered personalized support during office hours, assisting students in troubleshooting R scripts, interpreting statistical outputs, and applying appropriate analytical methods to solve real-world problems.
Evaluated and provided constructive feedback on student assignments, focusing on their ability to use R for exploratory data analysis, implement basic statistical tests, and create compelling data visualizations that support data-driven decision-making.
Skills: R, Python, Statistical Analysis, Data visualization, Storytelling

Sep 2024 - Present | Boston, MA

Infosys

Systems Analyst

Extracted time series data from Prometheus using SQL queries for ad-hoc analysis, enhancing data accessibility.
Utilized Python and R for statistical analysis and anomaly detection on time series data, identifying critical system issues. This led to a 40% reduction in unplanned downtime and a 25% improvement in mean time to resolution (MTTR).
Developed KPIs based on time series analysis, creating a monitoring framework that improved system efficiency by 20% and reduced operational costs by 15%.
Designed interactive Tableau dashboards using analyzed time series data, translating complex metrics into actionable insights. This facilitated a 35% increase in cross-team collaboration and supported data-driven decision-making at the executive level.
Tools: SQL, Python, R, Tableau

Dec 2021 - July 2023 | Hyderabad, India

Projects

                
Deep protein structure modeling
                  Prediction of tertiary structure of a protein from sequences of amino acids.
                
AccomplishmentsStandardized datasets like ProteinNet, derived from the CASP competition, are used for training models. Tools like TensorFlow-Keras, along with libraries such as NumPy and Matplotlib, help in building models for predicting protein structures from these datasets.

Explainable-AI-Healthcare
                  Improving accuracy of an explainable AI model’s prediction of patient outcomes by adding stacked generalization                
AccomplishmentsThe base learner, an Explainable Boosting Machine (EBM), provides feature importance scores, which are then used to train a meta learner for enhanced prediction. A useful tool for doctors to better understand and trust the model's outcomes.

Intrusion detection system
                  Exploring the NSL-KDD Dataset: A Comprehensive Analysis About Intrusion Detection System
                
AccomplishmentsBuilt and evaluated intrusion detection models using XGBoost and Logistic Regression, focusing on data cleaning, EDA, preprocessing, and model performance evaluation with metrics like accuracy and feature importance to improve network intrusion detection strategies.

Text2SQL LLM
                        Chatbot developed with Langchain and LlamaIndex for natural language to SQL processing, leveraging OpenAI's GPT-3.5 Turbo LLM engine
AccomplishmentsThe project aims to facilitate seamless interaction with IPEDS data, allowing users to generate SQL queries through natural language inputs, thereby enhancing accessibility to educational statistics and insights.

AI-Driven Marketing
                  Leveraging AI to enhance customer engagement and reduce churn                
AccomplishmentsCrafting personalized marketing campaigns across various channels (emails, SMS, ads) tailored to customer preferences and behaviors, and segmenting customers (New, Occasional, Regular, Lost) to predict churn and implement targeted retention strategies.

Risk Analysis of tech stocks
                  This analysis provides insights into stock performance, risk-return profiles of tech stocks from 2012 to 2023.                
AccomplishmentsRetrieved data via the yfinance library, calculated daily returns, visualized cumulative returns, computed volatility, performed correlation analysis, and calculated Value at Risk (VaR).

Insurance Fraud
                  This project aims to predict fraud in insurance policies using various machine learning models.                
AccomplishmentsIt focuses on understanding key factors linked to fraud and addressing class imbalance in the target variable. Multiple metrics, such as recall and AUC, are used to evaluate model performance and determine the best approach for predicting fraud effectively.

Online Course Dropouts
                This analysis utilized a LightGBM model to identify the primary indicators of student dropout in online courses.
AccomplishmentsKey findings revealed that consistent engagement, measured by the "Unique Days" feature, and interaction with video content were significant predictors of student retention, highlighting the importance of fostering a continuous learning experience that keeps students actively involved in their courses.

A/B testing
                  A/B Testing using Python to determine the most effective marketing campaigns.
AccomplishmentsThe goal is to analyze the impact of each campaign on weekly sales using libraries such as Pandas, Matplotlib, Seaborn, and Scipy to assess which campaign drives the highest sales across the randomly selected outlets.

YOLO Object Detection
                            The objective was to accurately identify and localize multiple objects within images
AccomplishmentsThe project involved several key steps, including data collection, annotation, and preprocessing of images to create a robust training dataset. I trained the YOLO model on this dataset, fine-tuning hyperparameters to optimize detection accuracy and speed.

Kafka-Spark
                            Real time processing and analytics of IoT data using Kafka and Spark
AccomplishmentsData from the sensors is streamed to Kafka, processed using PySpark with Spark Streaming, and then analyzed to derive insights such as average temperatures by state, total number of messages processed, and sensor distribution across states.