Rishikesh Ksheersagar

Rishikesh Ksheersagar

Masters in Data Science at University of Michigan
NLP Researcher at LIT Lab @ UMich
Previously - Data Science Manager at Mu Sigma Inc.

Data Scientist

Right after completing my Bachelor’s degree in Computer Engineering, I joined Mu Sigma as a Data Scientist in July 2019. I underwent 3 months of rigorous training where I gained some preliminary real-world experience and developed proficiency in Python, SQL, ML and Storyboarding.

World’s Largest Home Improvement Retailer

  • My first client engagement started in November 2019 where I was assisting the Financial Planning and Analysis (FP&A) team of a Fortune 25 US-based Home Improvement Retailer, in undergoing Digital Transformation.
  • We helped our clients develop an advanced tool to plan and forecast 7 Retail Metrics viz. Sales, Online Sales, Markups, Markdowns, Inventory, etc.
  • As my first assignment, I developed the entire codebase for planning the “Initial Markups” (IMU Metric) and wrote ~13,000 lines of Python code.
  • I was also responsible for maintaining the pipelines developed by us.
  • I got a chance to design and create 7 Tableau Dashboard from scratch - which were used to visualize the financial plans created by our tool and flag anomalies if any.
  • I was awarded my 1st SPOT Award due to this effort.
  • By December 2020, I was leading this team of 5 data scientists and managing the pipelines and dashboards for 4 of the 7 Retail Metrics.
  • Due to my consistent efforts and subject-matter, I recieved my 2nd SPOT Award.

Fortune 100 Telecom Clientele

  • Post ~2 years of gaining experience in the Retail domain, I moved to a team of 7 data scientists, as a lead, and worked with the Data Science and Data Engineering team of a US-based Telecom Giant.
  • We were partnering with the clients to develop and end-to-end framework to identify and tackle network impairments thus improving the “Network Service Experience”.
  • Initially, we started off with performing exhaustive EDA on previously un-explored datasets. Lack of Subject Matter Expertise really made this a challenging task.
  • From Cable Modem Registration Dataset, we identified over 14 network “events” which indicated a degraded network service on the consumer’s end for ~50M customers in near real-time.
  • We explored more datasets viz. Speed Tests, Modem Utilizations, Proactive-Network_maintenance, Noise ratios among others.
  • By the end of March 2022, we established an entire framework to identify any event indicating network failures with ~98% accuracy.
  • I was awarded my 3rd SPOT Award for my efforts here.