Divij Singh

Experience

• 

  HT Media Ltd

  Jun 2014 - Jun 2014

  Intern

  I interned as a writer for HT Campus. My responsibilities included writing about careers and the educational requirements associated with them, alerts about college admissions, and general interest articles.

• 

  Twyst.in

  Jun 2015 - Jun 2015

  Internship

  I interned at the company for 2 weeks. My work mostly entailed entering and sorting data on Excel sheets and getting location tags from Google.

• 

  Quatrro Business Support Services

  Jun 2018 - Aug 2018

  Intern

  Worked on building a website for internal use, using the Django web framework,including social media integration for tracking of user engagement.

• 

  Ashoka University

  Aug 2018 - Apr 2020

  Teaching Assistant and Cohort Leader

  • Teaching Assistant for the Introduction to Computer Programming course.• Corrected assignments, held and taught extra classes for students and assistedstudents with doubts.• Was a formal mentor for incoming students to Ashoka University in 2018 (10students) and 2019 (20 students).

• 

  Upgame

  Dec 2018 - Aug 2020

  Part Timer (While Studying)

  • Worked on experimentally mapping golf courses via CNNs, in order to automateone of the most time‐consuming tasks of the company.• Learning step involved processing roughly 3 gigabytes of images and associatedfiles.• Worked with clients on Data collection and analysis of players in golf tournaments,in order to give improved metrics to coaches.• Involved scraping several websites using Selenium and processing over 200,000rows of data per player.

• 

  The University of Aizu

  Jan 2021 - Mar 2022

  Graduate Teaching Assistant

  • Teaching Assistant for the Algorithms & Data Structures course, the IntegratedExercise for Software Course, and Operating Systems Course.• Corrected assignments, and assisted students with doubts.

• 

  Axelspace Corporation

  Apr 2022 - Jul 2022

  Intern

  Created a regression model and associated tools to model the power consumption of satellites and predict the satellite power at the next time interval, basedon component usage with 99.98% accuracy.

• 

  Rakuten Symphony

  Oct 2022 - Oct 2024

  Data Science Engineer

  • Improved existing Anomaly Detection pipelines which interact with APIs, Cassandra, and Kafka DBs. Converted pipelines from daily running to real-time detection.• Designed and created ML Pipelines for Automatic Root Cause Analysis of Network Metrics, to automatically create detailed reports of Kubernetes cloud issues when running internal use Spark jobs.• Researched Network Alarm Compression using pattern mining for rule creation, and successfully implemented the ACOR method described by Philippe Fournier-Viger to successfully compress alarms by 40%.• Researched methods to impute time series data, successfully implemented the imputation by decomposition method, then further improved it to more closely represent the missing data.• Designed and successfully implemented Kale pipelines to ingest time series data, extract descriptive information, impute missing values and re-write data. Pipelines are being used as a core part of KPI monitoring for the platform.• Designed and successfully implemented MLFlow tracking for model testing Show less