Summary
Overview
Work History
Education
Skills
Awards & Recognitions
Timeline
Generic
SUBBARAJU I

SUBBARAJU I

Asst Delivery Manager

Summary

Organized techo-functional Delivery Manager and an Individual Contributor in Machine Learning, Deep Learning advanced Analytics and Platform Engineering to deploy the models with over 13 years of experience in Banking, Procurement and Health Insurance domains with a proven record of delivering high performance results.

Overview

14
14
years of professional experience

Work History

Asst Delivery Manager - DS & Engg

Carelon Global Solutions
11.2021 - Current
  • Managing a team of 60 associates, spread across data science, QA, platform engineering, DevOps and UI to deliver end to end solutions
  • Developed and managed the team to come up with the best prompt with the help of prompt engineering and deployed the LLM models in production for real time use
  • Experience exploring the open source LLM's – Mistral 7B, Falcon 40B, OpenLLaMA models to bench mark the results with ChatGPT
  • Developed tesseract, textract and google vision OCR solutions to extract the text out of the documents
  • Built the NLP model based on the text extracted from the OCR to help the clinicians review the pre-auth cases efficiently with the help of coder model, which is based on BERT
  • Improved the existing NLP solution with the help of commercial open AI and Open source LLM models to extract the relevant evidences for the Inpatient policies
  • Worked with the team of data scientists, data engineers, UI and QA testers to make sure the solution was delivered in time and without any issues
  • Assisted and part of multiple PoC's to enhance the existing solution
  • Spread headed the innovation council to improve the solutions with the help of LLM models

Principal Data Scientist

TVS Credit Services
06.2021 - 10.2022
  • Developed an application fraud score model to prevent fraud
  • Developed the classification models to help the collections team prioritize their calling based on the customer's likelihood to pay
  • Helped the team to improve their collections by 66% by targeting the right customers to call at the right time
  • Developed various trigger based systems to identify fraud
  • Played a key role in building an effective frame work for fraud
  • Utilized unsupervised and neural network architectures to build robust models for fraud mitigation.

Manager - Data Sciences

Synchrony International
01.2017 - 06.2019
  • Developed strategies to minimize the transactional fraud in the retail sector across the US
  • Built an Autoencoders network for anomaly detection to identify the fraudulent transactions
  • Developed the Random Forest and XG Boost techniques to identify the true name fraud using Python
  • Developed an Early Warning System (EWS) and dashboard to identify the Fraud, which eventually saved $2.5MM and also helped in identifying the production errors
  • Developed a fallback strategy to reduce the fraud in keyed transactions, which saved ~$200K
  • Designed and implemented a strategy for High Velocity Customers, which eventually helped the business to increase the sales by $8MM per week
  • Helped the business to migrate from FICO to SAS score by proving an extensive data deep dives and performance metrics
  • Worked with decision trees to identify important variables to use in developing the strategies.

Senior Risk Analyst

Financial Bank USA Support Services Pvt. Ltd
05.2012 - 12.2016
  • Model Development Projects
  • Developed 30 DPD to Charge-off model to predict the probability of 30 days past due accounts getting charged-off in next 12 months and validated the model
  • Built a logistic regression model for getting the probability of a 30 day delinquent account going on to 60 day next month, validated the model and compared the performance with the existing model
  • Developed a strategy using decision trees for early Credit Line Increase (CLI) to select the population at month 3 instead of month 6
  • Collections Projects
  • Worked on the ‘Reinstatement' project to re-issue the charging privileges back to the customers
  • Worked on bubble reports in various buckets to identify various factors leading to an increase in the corresponding bucket openings
  • Worked on developing, validating and comparing the dialer metrics like call penetration, connects, contacts, promises etc.

Analyst

Technopal Pvt Ltd
07.2010 - 05.2012
  • Worked on a project to identify the accounts used Over Draft (OD) facility and track their performance
  • Identify and select the population eligible for the limit increase
  • Developed macros for data cleaning and processing

Education

B.Tech - Electronics & Communications

Sri Muthukumarn Institute of Technology
01.2002 - 2006.04

MBA - Marketing & Operations

NIILM School of Business (NSB)
01.2008 - 2010.04

Skills

    Machine Learning, Deep Learning, NLP, Computer Vision

undefined

Awards & Recognitions

  • Multiple IMPACT awards for displaying the leadership and contribution towards deliverables
  • Received an ‘Outstanding Star' award for developing the EWS & Auto Encoders
  • Received a ‘Bright Star' award for collaboration in developing the Random forest model to identify the fraud
  • Received an ‘Outstanding Star' award for flawless execution during the Peak Season

Timeline

Asst Delivery Manager - DS & Engg

Carelon Global Solutions
11.2021 - Current

Principal Data Scientist

TVS Credit Services
06.2021 - 10.2022

Manager - Data Sciences

Synchrony International
01.2017 - 06.2019

Senior Risk Analyst

Financial Bank USA Support Services Pvt. Ltd
05.2012 - 12.2016

Analyst

Technopal Pvt Ltd
07.2010 - 05.2012

MBA - Marketing & Operations

NIILM School of Business (NSB)
01.2008 - 2010.04

B.Tech - Electronics & Communications

Sri Muthukumarn Institute of Technology
01.2002 - 2006.04

Similar Profiles

CREATE PROFILE
SUBBARAJU IAsst Delivery Manager