Vamshidhar Reddy Bollampally

• Developed machine learning models to track employee performance, predicting weekly and monthly productivity trends using time-series forecasting and regression methods
• Built sentiment analysis models to analyze employee feedback, using language models (LLM) for more accurate feedback and sentiment interpretation
• Created and deployed interactive dashboards with Streamlit & Power BI, allowing managers to easily track employee progress and task completion
• Improved machine learning models to make predictions on employee performance more accurate, using clustering and classification techniques
• Integrate the fake news detection model into the university portal by developing RESTful APIs using Flask or FastAPI
• Monitor the model’s performance, analyze errors or false positives/negatives, and retrain the model as new data becomes available
• Perform data cleaning and text preprocessing, including removing irrelevant content, tokenizing text, and converting it to a machine-readable format
• Design and train machine learning models for classifying news as fake or real by Logistic Regression, Naive Bayes

• Designed and implemented scalable data pipelines using Python, Spark, and AWS services, optimizing ETL workflows and reducing data processing times by 30%
• Automated data ingestion and transformation processes using AWS Glue, Apache Kafka, and Lambda, improving data reliability and reducing manual intervention
• Migrated on-premise databases and ETL workflows to AWS, implementing optimized cloud-based architectures using Redshift, S3, and DynamoDB for scalability and performance
• Developed real-time data streaming solutions with AWS Kinesis, Apache Flink, and Lambda, enabling low-latency data processing for critical business applications
• Optimized data warehouse performance using SQL, Snowflake, and BigQuery, ensuring efficient querying, improved data integrity, and enhanced reporting capabilities

Build Chatbot using Neural Network:

● Built a chatbot using Natural Language Processing (NLP) techniques such as tokenization, stemming, and bag-of-words for text preprocessing.

● Implemented a neural network model using TensorFlow and TFLearn for intent classification and response generation.

● Leveraged libraries such as TensorFlow, TFLearn, NLTK, NumPy, and JSON for building and training the chatbot.

● Managed the entire ML lifecycle, including data preprocessing, model training, evaluation, and deployment.

Face recognition:

● Developed a live face detection system using OpenCV, capturing and processing video streams from webcam input to enable real-time analysis.

● Utilized Haar Cascade Classifiers with detectMultiScale to identify faces, tuning parameters (scaleFactor, minNeighbors) to optimize accuracy and reduce false positives.

● Designed dynamic bounding boxes around detected faces using OpenCV’s drawing utilities, enhancing user interaction and visual feedback.

● Balanced detection efficiency and speed by adjusting frame resolution, grayscale conversion, and classifier parameters for smooth real-time execution.

Spam SMS Classification:

● Created a machine learning model to sort SMS messages into "Spam" or "Ham" categories, using different methods like Naive Bayes and Random Forest, achieving a high accuracy score.

● Improved model performance by adding features like word count and numerical values and balanced the data to ensure better results.

● Cleaned and prepared text data by removing unwanted characters, converting text to lowercase, and simplifying words for better model understanding.

● Tested the model’s performance and improved it by using techniques like ensemble methods, making it more accurate in detecting spam messages.