Prashant Raut Email & Phone Number

Pune, Maharashtra, India

Pune, Maharashtra, India

Accountable for overseeing end-to-end development stories and addressing critical production bugs throughout the software development lifecycle.Actively participated in the process of gathering project requirements.Implemented the method for Generating training data to support the development of the classification model.Developed a Convolutional Neural Network (CNN) model for checkbox classification, determining whether each checkbox is selected or not.Improved the model… Show more Accountable for overseeing end-to-end development stories and addressing critical production bugs throughout the software development lifecycle.Actively participated in the process of gathering project requirements.Implemented the method for Generating training data to support the development of the classification model.Developed a Convolutional Neural Network (CNN) model for checkbox classification, determining whether each checkbox is selected or not.Improved the model Accuracy.Utilized this model to replace Textract Analyze API with the more cost-effective Textract Document Detect API. This strategic change not only led to substantial cost savings but also elevated the solution's accuracy.Designed and implemented a Debug Shift tool to visualize bounding box plots on images.Developed and deployed a mechanism to transfer template form data from AWS textract response to a dedicated database table.Developed a Named Entity Recognition (NER) model to identify specific textual information within forms accurately.The NER model is built on the Transformer architecture, utilizing the pre-trained Spacy en_core_web_trf model.Deployed the NER model using FastAPI on an EC2 instance.Devised and implemented a targeted text extraction approach, utilizing AWS Textract, to retrieve data from images and PDFs for the NER.Established FastAPI endpoints to employ sophisticated NLP techniques for parsing unstructured address strings into distinct address components.Engaged in optimizing code and enhancing application performance.Engaged in the deployment and build process.Proficiently handled debugging processes and promptly resolved issues to ensure smooth progress. Show less

Pune, Maharashtra, India

I have undergone comprehensive training in Data Science, Machine Learning, Deep Learning, and AWS AI services, establishing a strong foundation in these cutting-edge technologies.During my training, I acquired valuable skills in deploying models using Flask and FastAPI in the AWS EC2 instance, Heroku, etc.Developed a Proof of Concept (PoC) for model deployment, gaining hands-on experience in the process.Actively participated in requirement-gathering processes.Successfully… Show more I have undergone comprehensive training in Data Science, Machine Learning, Deep Learning, and AWS AI services, establishing a strong foundation in these cutting-edge technologies.During my training, I acquired valuable skills in deploying models using Flask and FastAPI in the AWS EC2 instance, Heroku, etc.Developed a Proof of Concept (PoC) for model deployment, gaining hands-on experience in the process.Actively participated in requirement-gathering processes.Successfully implemented checkbox localization methodology on images.Contributed to image preprocessing techniques as part of my responsibilities.Led the design and development of the backend API for the application.Effectively implemented image orientation detection module to rectify image orientation issues.Since orientation detection is done with Tesseract, there's no need for an additional call to Textract, resulting in reduced API costs.Employed OpenCV's template matching method to determine whether image registration was necessary, and subsequently implemented the image processing technique. This approach aligns input images to template forms, enhancing overall consistency.Devised a dynamic table extraction algorithm for extracting tabular structures from forms, including borderless tables. This solution significantly enhanced table detection capabilities.Proficiently handled debugging processes and promptly resolved issues to ensure smooth progress.Engaged in optimizing code and enhancing application performance.Engaged in the deployment and build process.Implemented the new functionality as per product requirements and provided the solutions. Show less

Pune, Maharashtra, India

Led the end-to-end development of a dynamic Real-Time Twitter Sentiment Analysis and Entity Recognition application. This innovative solution enables the extraction of real-time tweets through specified keywords, subsequently determining the sentiment and identifying entities within each tweet. Leveraging the Twitter API, Orchestrated the collection of tweets based on designated keywords, followed by sentiment analysis using Amazon Comprehend. The results were then presented through a real-time… Show more Led the end-to-end development of a dynamic Real-Time Twitter Sentiment Analysis and Entity Recognition application. This innovative solution enables the extraction of real-time tweets through specified keywords, subsequently determining the sentiment and identifying entities within each tweet. Leveraging the Twitter API, Orchestrated the collection of tweets based on designated keywords, followed by sentiment analysis using Amazon Comprehend. The results were then presented through a real-time dashboard crafted with Python's Plotly Library, offering intuitive data visualization.My responsibilities encompassed several pivotal aspects of this project:Twitter Data Extraction: Implemented the method for retrieval of tweets from Twitter's API, tailoring the results based on specific keywords, thus ensuring relevant data collection.User Interface Development: Created a user-friendly interface using Flask, allowing seamless extraction of tweets based on user-defined keywords.Sentiment Analysis Integration: By seamlessly integrating the Amazon Comprehend Sentiment Analysis API, enabled the system to accurately evaluate the sentiment of each tweet, contributing to a comprehensive understanding of public perception.Entity Recognition Implementation: Harnessed the capabilities of the Amazon Comprehend Entity Recognition API, allowing the system to recognize and extract named entities within the tweets, adding depth to the extracted information.Lambda Code for Amazon Comprehend: To enhance efficiency, devised optimized Lambda code to facilitate the integration of Amazon Comprehend APIs.Real-Time Dashboard Creation: Crafted an interactive real-time dashboard utilizing Python's Plotly Library. This visualization mechanism succinctly conveyed the sentiment analysis and identified entities, granting users a swift grasp of the data's insights.Led the creation of a dynamic Python-Flask tool using AWS for real-time Twitter sentiment & entity analysis. Show less

Pune, Maharashtra, India

Bachelor Of Engineering - Be, Computer Science, 8.24/10

Education record