Dhruv Tyagi

Data Science and Machine Learning student with internship experience at Travelers Insurance, specializing in NLP and large language models for automating unit test generation. Developed an Isolation Forest-based anomaly detection system for KPI time-series data. Proficient in Python, SQL, Databricks, PySpark, and MLflow, with expertise in building scalable pipelines and deploying ML models. Certified AWS Cloud Practitioner with a focus on predictive modeling and cloud-based analytics, dedicated to leveraging AI/ML for data-driven decision making.

Detecting Parkinson’s Disease with XGBoost

• Built a classification model to detect Parkinson’s disease from medical voice recordings.
• Applied signal preprocessing and trained with XGBoost, tuning hyperparameters to optimize performance.
• Achieved 93% accuracy, demonstrating potential for early clinical screening applications.

Handwritten Digit Recognition (CNN on MNIST)

• Designed and implemented a Convolutional Neural Network (CNN) in TensorFlow/Keras for digit recognition.
• Applied data augmentation to reduce overfitting and improve generalization.
• Achieved 89% accuracy, validating the effectiveness of deep learning for image classification.

Historical Weather Data Analysis for Air Quality Prediction

• Developed a predictive model using weather, urban activity, and traffic datasets to forecast air quality.
• Conducted feature engineering and used ensemble models to improve accuracy by 78% over baseline.
• Showcased ability to integrate multiple real-world datasets for environmental health analytics.

Music Genre Classification

• Built an ML model to classify songs into genres using Librosa for feature extraction and Scikit-learn for classification.
• Trained on audio datasets and achieved 75% accuracy, demonstrating skills in signal processing and supervised learning.

• AWS Certified Cloud Practitioner
• AWS Certified AI Practitioner
• CS50’s Introduction to Artificial Intelligence with Python — edX