Ayush Singh

Led development and optimization of predictive models for retail clients using logistic regression and random forest, resulting in a 15% increase in customer retention. Improved forecasting accuracy of a $500 million retail portfolio by implementing ARIMA and Prophet time-series models, reducing forecasting errors by 20%. Deployed scalable machine learning models including gradient boosting and XGBoost, reducing inference time by 30%. Collaborated with cross-functional teams to translate business problems to data-driven solutions, leveraging K-means clustering and decision trees, driving a 10% revenue increase through optimized marketing strategies.

Designed and deployed real-time and batch data pipelines on Google Cloud Platform for large-scale messaging analytics. Engineered streaming pipelines using Pub/Sub for event-driven data ingestion and Apache Beam (Dataflow) for data transformations. Optimized storage and querying with BigQuery for analytics and reporting. Applied Generative AI, including LLMs like BERT and Gemini, with prompt engineering to enhance automated text classification and fraud detection. Developed AI/ML models for anomaly detection and pattern recognition, improving fraud identification efficiency. Utilized clustering algorithms and NLP techniques for text classification tasks.

Developed REST API to extract data from multiple tables within editable PDFs, consolidating data into Excel files for processing. Performed detailed data analysis identifying trends and insights supporting strategic decision-making. Implemented data cleaning and transformation techniques reducing inconsistencies by 30%.