Sushil S

Email: [email protected]
Mobile: (+91) 9178457701

PROFESSIONAL OVERVIEW

Data Science professional with 14 years of experience in leveraging data-driven insights and machine
learning techniques to address complex business problems across Banking, Telecom, Energy, and Retail
sectors. Proficient in harnessing the power of Large Language Models (LLMs) for developing innovative
solutions and driving strategic decision-making. Experienced in team management, strategic planning,
and solution architecture.

Key Skills:
 Statistical Modeling, Machine Learning
 Linear Regression, Logistic Regression, Decision Tree, Random Forest, Cluster Analysis, Deep
Neural Network, Convolution Neural Network, Recurrent Neural Network, Time Series Analysis
 Natural Language Processing (NLP), Computer Vision
 Python, PyTorch, PySpark, R programming language, SAS, SQL
 Transformers, BERT, Large Language Model (LLMs), RAG, Langchain, Open AI, Private LLM,
Vector Database, Graph Database, Prompt Engineering
 LLM fine tuning, Parameter Efficient Fine-tuning (PEFT), LoRA training, LLM serving

PROFESSIONAL EXPERIENCE

Group Solution Leader - Data Science

Softream It Services
February 2023 – Present

 Design, develop and implement Generative AI based applications to solve diversified business
use-cases.
 Fine-tuning of large language models on custom data and for specific task.
 Lead the development of Enterprise Information Search Tool using private LLM that allows users
to engage in natural language conversation directly with the documents of their enterprise
knowledge base.
 Lead the development of Brane Dynamic product, which facilitates interaction with SQL
databases using natural language. This product is designed to provide an interactive and intuitive
interface for Enterprise users to query their relational databases in English and receive responses
in text, tables or charts based on the nature of the query.
 Optimize LLM inferencing pipeline for private LLM to achieve higher throughput, lower latency
and enabled streaming output for reducing apparent latency to the end users.
 Implementing user friendly features like autosuggestion, query recommendation, and a caching
layer to enhance user experience with Gen AI products.
 Setup and grow the team of Data Scientists.
 Provide technical guidance and mentorship to junior team members.

Page 1 of 4
Lead - Data Science
Bajaj Finance Limited, Pune
March 2021 – November 2022

 Working with Business and Product management teams on designing, building and deploying
data products using customer data and alternate data.
 Developed financial products using alternate data like Reverse Bank Statement, Reverse Credit
Card Statement & Wallet Transactions Report that helps analyze customer’s spending and risk
behavior.
 Developed customer face match solution using ML techniques for identifying wrongly mapped
mandate of customers for loan repayment.
 Developed CAPTCHA solver using ML for scrapping alternate data from publicly available
websites like EPFO.
 Development of Propensity model to drive cross-sell of financial products to existing Bajaj
customers.
 Mentoring of junior team members.

AI Architect
Wipro Limited, Bengaluru
September 2020 – March 2021

 Worked on developing Semantic Search Engine for one of the world’s largest petrochemical
manufacturer for their internal scientific documents using NLP techniques and Knowledge Graph.

 Leveraged advanced NLP techniques like Named Entity Recognition (NER), keyword extraction,
and sentence embeddings for improved search accuracy.

 Created a Graph Database using JanusGraph to interlink research topics, regulatory guidelines,
and supporting documents.

 Implemented question-answering systems using transformers for querying critical information

from regulatory documents.

Data Science Architect

Perfios Software Solutions Private Limited, Bengaluru
December 2018 – August 2020

 Developed the Risk Scorecard which helps financial institutions (Bank and NBFC) to assess
credit risk of the customers who are applying for loan based on the transaction history from their
bank statement
 Developed a solution using fuzzy logic which helps in the identification of window dressing with
98% accuracy in the bank statements of the SME organizations applying for business loan.
 Developed Machine Learning based solution to correct the error/mistakes in the output of
scanned bank statement produced by OCR using NLP and fuzzy logic and reduced manual
intervention by 40%
 Identification of fraud/tampered pdf of bank statement using ML algorithm.
 Working on the AI based KYC product where customers’ real time photo is verified with his
photo in the ID card, face liveness check is performed using deep learning models and his details
from the ID card is extracted using OCR technology.

Page 2 of 4
Assistant Manager – Advanced Analytics
Tata Consultancy Services, Kolkata
November 2017 – December 2018

 Developed Machine Learning based solution to harmonize external sales data (obtained from
Nielsen) to the internal Product Master of one of the world’s largest alcoholic beverage company
for their various markets using Microsoft Azure ML Studio
 Developed Sales forecasting dashboard using Power BI for different categories of Product for the
same alcoholic beverage company for their different markets in Europe

Data Scientist
EdgeVerve Systems Limited (an Infosys’ Company), Bengaluru
October 2015 – November 2017

 Working with Engineering, and Product Management teams on designing, building and deploying
data analysis systems/products for Banking industry.
 Working on Next Best Offer strategy for Banks that will predict the next best offering of banking
products for its customers to improve customer loyalty and to boost cross-sell of products.
 Developing machine learning techniques in the area of text classification and sentiment analysis.

Business Analyst
Positive Integers Private Limited, Chennai
June 2014 – October 2015

 Responsible for delivering high quality, in-depth insights on Customer Churn Management,
Mobile Data Penetration, Mobile Data Elasticity, High Value Customer Acquisition for Positive
Integers clients
 Establishing scalable, efficient, and automated processes/codes for large scale data analysis

Engineer
Sterlite Industries India Limited (a subsidiary of Vedanta Resources), Tuticorin
August 2009 – April 2013

 Data driven analysis and improvement of Power Plant efficiency

TRAINING AND CERTIFICATION

 SAS Certified Base Programmer for SAS 9

Page 3 of 4
 SAS Certified Statistical Business Analyst Using SAS 9: Regression and Modeling

Page 4 of 4