Ravi Putcha developed three end-to-end data science features in the nikbearbrown/INFO_7390_Art_and_Science_of_Data repository, focusing on customer segmentation, language model fine-tuning, and automated travel itinerary generation. He built a causal inference notebook suite using Python and Jupyter, enabling data-driven segmentation analysis. For language models, he implemented few-shot learning with LoRA parameter-efficient fine-tuning, applying PyTorch and Hugging Face Transformers to sentiment analysis tasks. Additionally, he engineered a full pipeline for itinerary generation, integrating web scraping, AWS S3, Snowflake, and a chat interface. The work demonstrated depth in data engineering, machine learning, and backend development, delivering robust, production-like solutions.
April 2025 — Key features delivered across nikbearbrown/INFO_7390_Art_and_Science_of_Data: 1) Causal Inference Notebook Suite for Customer Segmentation; 2) Few-shot Learning Notebook for Language Models with LoRA; 3) AI-Generated Travel Itinerary End-to-End Pipeline. These efforts deliver end-to-end data science capabilities: from data loading and causal analysis to parameter-efficient model fine-tuning and a production-like data pipeline with web scraping, embeddings, and a chat interface. Business value includes improved segmentation decision support, faster experimentation with LoRA-based fine-tuning that reduces compute, and scalable, automated itinerary generation.
April 2025 — Key features delivered across nikbearbrown/INFO_7390_Art_and_Science_of_Data: 1) Causal Inference Notebook Suite for Customer Segmentation; 2) Few-shot Learning Notebook for Language Models with LoRA; 3) AI-Generated Travel Itinerary End-to-End Pipeline. These efforts deliver end-to-end data science capabilities: from data loading and causal analysis to parameter-efficient model fine-tuning and a production-like data pipeline with web scraping, embeddings, and a chat interface. Business value includes improved segmentation decision support, faster experimentation with LoRA-based fine-tuning that reduces compute, and scalable, automated itinerary generation.
Overview of all repositories you've contributed to across your timeline