This POC/ boilerplate API is part of a larger system implementing data-warehousing that was required to normalize data coming from various sources and of different types. The system parsed all the incoming data to common JSON format, before ingestinng to the ware-house. We also parsed PDFs, xmls, html and excel/ csv files along with other databases.

Data-Warehouse related project for fetching, parsing and mapping commercial invoices with relevant delivery-challan & purchase-order. Parsing the Purchase Orders, Delivery Challans & preparing mappings.

Top 3 Key Learning:

• Setting up boilerplates for flask project
• POc development with Flask templates
• Data-warehousing

Build Chatbot application for Telecom Industry client

Using Technologies Stack: Python, RASA, Tensorflow, Spacy, Flask etc.

Bot was responsible for providing informtion to users and carrying out different transaction on behalf of verified users.

Analyzing Biomedical Images for assisting doctors with pre-examination Biopsy Information for Cancer Treatment.

This application is a sample of in-house built product Learning Management System with modern cutting edge technologies to ease the learning process. Implemented core backend engine and supporting apps for handling accounts, courses, assignments, discussions and quizzes.

A contextual chatbot system to assist/ replace human dependency in Telecom customer care up to 90 by using conversational AI technologies. We used NLP techniques (like intent classification, response retrieval entity extraction for better and near real time user experience with chatbots. Used Rasa along with other custom build models for dialogue management language understanding. Built REST api for interaction with frontend app.

A Natural Language Processing application to parse legal contract terms, analyze data, classification into multiple predefined categories and detecting key entities using NERs.

AI enabled solution for biopsy quality determination in real-time to simplify digital pathology workflow and reduces the time clinicians spend assessing a slide by providing the key tissue information and sample metrics, all calculated from a whole slide image.