Kriti Agrawal Email and Phone Number

• Worked with an American Multinational brewing Company to revamp their data driven guided selling solution using advanced AutoML techniques to identify potential new points of distribution for their products• Developed a 5-year forward forecast for Beer/RTD products using Linear ARIMA and Prophet models, providing critical insights at the provincial level for strategic planning and inventory management• Developed an innovation script to identify optimal placements for new products… Show more • Worked with an American Multinational brewing Company to revamp their data driven guided selling solution using advanced AutoML techniques to identify potential new points of distribution for their products• Developed a 5-year forward forecast for Beer/RTD products using Linear ARIMA and Prophet models, providing critical insights at the provincial level for strategic planning and inventory management• Developed an innovation script to identify optimal placements for new products, driving market entry strategies for novel offerings• Contributed to business development by creating a demo sales monitoring dashboard in Power BI for a potential client, showcasing data visualization and analysis capabilities Show less

• Extensively worked on Covid Misinformation spread via tweets, paper in review at an esteemed journal• Fine-tuned frameworks based on unsupervised embeddings generated from Word2vec, Glove, Fasttext, BERT-based architecture integrated with deep learning models for classification• Used LDA to identify the most relevant topics in the dataset, and performed category classification on them using BERT Embeddings• Used Plutchik’s transformer model to detect underlying emotions in the… Show more • Extensively worked on Covid Misinformation spread via tweets, paper in review at an esteemed journal• Fine-tuned frameworks based on unsupervised embeddings generated from Word2vec, Glove, Fasttext, BERT-based architecture integrated with deep learning models for classification• Used LDA to identify the most relevant topics in the dataset, and performed category classification on them using BERT Embeddings• Used Plutchik’s transformer model to detect underlying emotions in the tweets Show less

• Conducted extensive research on bounding box detector algorithms to identify the most effective methods for detecting SKUs for Pepsi products in images from retail shops in Bangladesh• Evaluated the existing YOLOV5 algorithm and identified key improvement areas where it was falling short• Retrained the YOLOV7 algorithm while customizing the hyperparameters to better suit the unique characteristics of the dataset, resulting in a 10% increase in accuracy compared to the previous… Show more • Conducted extensive research on bounding box detector algorithms to identify the most effective methods for detecting SKUs for Pepsi products in images from retail shops in Bangladesh• Evaluated the existing YOLOV5 algorithm and identified key improvement areas where it was falling short• Retrained the YOLOV7 algorithm while customizing the hyperparameters to better suit the unique characteristics of the dataset, resulting in a 10% increase in accuracy compared to the previous algorithm• Utilized topic modeling techniques to cluster large amount of unstructured Hindi text data, consisting of posts with no clear structure into distinct topics, resulting in 8-10 well-defined clusters and created a structured dataset that can be used for future posts and analysis• The labeled dataset enabled the client to better understand their audience and tailor their content strategy accordingly, resulting in a 15% increase in engagement Show less

• Engineered an NLP pipeline using Pandas, NLTK, and SpaCy to extract and quantify skill proficiency from 100,000 job descriptions• Conducted extensive research and analysis of various skill proficiency data sources, leading to the compilation of a comprehensive database of skills• Extracted proficiency for different skills in various jobs from the scraped data using NLP techniques, quantified and standardized it, resulting in a 25% improvement in accuracy in skill matching

• Successfully qualified for Microsoft’s mentorship program and built a Microsoft Teams Clone with video and chat functionalities using Agile Methodology, resulting in a 95% project completion rate• Effectively integrated the application with Twilio-based APIs to deliver text and video call functionality, ensuring smooth and uninterrupted communication between users• Built the Teams Clone using HTML, CSS, and JS, utilizing the Django framework for web development, resulting in a… Show more • Successfully qualified for Microsoft’s mentorship program and built a Microsoft Teams Clone with video and chat functionalities using Agile Methodology, resulting in a 95% project completion rate• Effectively integrated the application with Twilio-based APIs to deliver text and video call functionality, ensuring smooth and uninterrupted communication between users• Built the Teams Clone using HTML, CSS, and JS, utilizing the Django framework for web development, resulting in a user-friendly and intuitive interface• Hosted the Teams Clone on Heroku, ensuring reliable and efficient access for users Show less