Jul 2018 – May 2021 · 2 yrs 10 mos · Gurgaon, India

• DQIML : Data Quality Improvement using Machine Learning. Web portal having ML prediction models for predicting Responsible Party, Severity, Reason For Outage for a resolved ticket. Tech : Python, Django, MongoDB, NLP, Supervised ML, FastAPI.
• NLP Classification - Predict severity and responsible party based on closure comments, router type, assumed severity, etc.
• Prediction Service - Async API to access prediction models.
• Web Portal (Django) - Web portal for user to consume prediction service.
• SOBOT : Service Operation BOT. It troubleshoots the new incidents with > 90% success rate. Tech : Java, Python, ML, Mysql. Keys tasks delivered:
• Asynchronous API - Implemented async API using python's FastAPI for fast parallel troubleshooting of new tickets.
• Top 3 Problem Families - Recommending top 3 problem families based on previous history.
• Single code base - Merging all SOBOT customer specific code bases to run through common code base and devops.
• Improving Bot Performance - upgraded library versions to run on latest firefox, removed redundant DB and function calls, applied intelligent waits , etc.

Jun 2017 – Jul 2017 · 1 mo · Ghāziābād Area, India

• Made DCCC RADAR SENSOR SIMULATOR in JAVA with gui using which data is send and received through UDP connection.