• Created the foundational dataset for a change detection model using QGIS software

• Processed and accurately labeled 98 Cartosat-2 and 295 Cartosat-3 high-resolution image pairs

• Produced high-quality labeled data that enabled effective training and validation of a semi-supervised change detection algorithm

• Contributed directly to improved model accuracy in identifying temporal changes

• Gained hands-on experience applying machine learning and data visualization techniques to real-world business problems during a data science internship at Ingrade

• Built and fine-tuned a BERT-based sentiment analysis model to extract actionable insights from textual data

• Used Tableau to visualize complex financial datasets and develop interactive dashboards highlighting key performance indicators

• Authored technical blogs to support knowledge sharing, with three articles featured on Ingrade's official LinkedIn page

• Designed and implemented a high-accuracy resume extraction pipeline using Large Language Models (LLMs) and FastAPI

• Reduced information retrieval time by 66.67% through optimized processing and automation

• Engineered a highly precise prompt for the Gemini model, improving resume parsing and feature extraction accuracy from 60% to over 95%

• Integrated Mixpanel and PostHog to strengthen website analytics and user behavior tracking