Developed Chatbot for the Customers & Vendors who are looking frequent questionnaire and connect to support teams for queries globally.With Chatbot 9million business generated by resolving queries within shot span of time and improved quality work force in teams

From Over 50k servers globally logs were generated in huge amount of size in servers. To identify RCA ( Root Cause Analysis) and monitoring used ELk based approach for to identify different scenarios as process oriented

Developed an automated tool to categorize text data. Documents will be updated periodically and tool should be able to categorize the text and find new categorization. Extracting data from cluster and translating all languages into English and summarized each document to segregated data by categorizing and grouping of words, sentences

Developed Chatbot for sales and marketing team using Rasa-NLU, NLP to give auto responsive feedbacks raised by customers. To develop this bot, collected data from various sources to store large corpus. Applied NLP techniques, PySpark to structured data and extract information, Identify topic entities from the corpus. Created trainer and chat servers to centralized business units.

Designed and Implemented monitor based performance dashboard using ELK(Elasticsearch) and Kafka(PyKafka).Used Kafka to collect server logs and cpu logs data from different sources as a producer and transfer to relevant buckets were topics created to store and process data to consumer end as a realtime streaming showcase in Kibana dashboard. Applied Time Series analysis algorithm to predict and trigger server faults.

To analyze the data captured from machine processors memory modules located at various regions. Using this data define some KPIs to predict performance of machines. Develop metrics and indices to provide users with a layer of structured analytics on top of a high volume and unstructured data set.

Worked with in-house research medical team to gather data and inputs for building and predictive text search engine.Developed a standalone application to pull out medical articles, clean the data and cluster into sub-groups .Using LSTM & DNN applied to get classified , predicted and optimized number of drugs for generic drugs.