Rajat Thakur

Data Scientist

Bengaluru, Karnataka, India8 yrs 8 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

8+ years of experience in AI/ML.
Expert in developing predictive models for fintech applications.
Strong collaboration with cross-functional teams.

Stackforce AI infers this person is a Fintech Data Scientist with expertise in AI/ML and predictive modeling.

Contact

Skills

Core Skills

Data ScienceMachine Learning

Other Skills

ARIMAAlgorithmsApache SparkArtificial Neural NetworksBig DataBusiness Intelligence (BI)C++CatBoostClassificationCollaborative FilteringCredit Risk Modeling ToolsData MiningData ModelingData ProcessingData Structures

About

Experience

Flipkart

2 roles

Senior Data Scientist

Oct 2024 – Present · 1 yr 5 mos · Hybrid

Data Scientist

Mar 2020 – Sep 2024 · 4 yrs 6 mos · Hybrid

Key contributor to Fintech Data Science team for Credit Risk, Transaction Fraud, Collection Recovery & Propensity Modeling.
Developed Transaction Fraud models using CatBoost for Pay Later/EMI.
Developed Propensity models for Pay Later, Co-branded Card & Personal Loan to drive marketing campaigns and increase funnel conversions
Worked on Encoder-Decoder consisting of LSTM/GRU built with PyTorch. Also tried out Seq2Seq machine translation.
Created a Collection Scorecard to predict due repayment probability. Reduced NPA & Collection cost. Published a paper in Flipkart’s Data Science conference Convergence 2022.
Developed Claim Risk models for Mobile Protection & Life Insurance.
Experimented with Bayesian Neural Networks (BNNs) with TensorFlow for quantifying prediction uncertainty.
Performed large-scale extensive Feature engineering using HQL, PySpark.
Collaborate actively with cross-functional stakeholders from Business, Risk, Marketing, and Product.

Credit Risk Modeling ToolsCatBoostPyTorchTensorFlowHQLPySpark+2

Airtel x labs

Associate Data Scientist

Jun 2017 – Feb 2020 · 2 yrs 8 mos · Gurugram, Haryana, India

Implemented Genre Prediction consisting of Data Processing, Feature Engineering, Frequent Pattern Mining & Multi-Label Classification using PySpark for ~100 Million customers.
Implemented Content Recommendation Engine using Collaborative Filtering.
Time Series Forecasting of 4G Data Consumption using ARIMA models
Central Data Warehouse: Implemented Spark Jobs for ingesting raw data Records (order of Terabytes/Day) into HBase & Integrated Apache Phoenix to fetch data from HBase via SQL queries.

PySparkCollaborative FilteringARIMAHBaseData Science

Ibm india private limited

Extreme Blue Program, Summer Internship

May 2016 – Jul 2016 · 2 mos · Bengaluru, Karnataka, India

Implemented an efficient approach for Disaster Recovery in IBM's spectrum scale-Asynchronous Disaster Recovery (ADR).
Precise estimation of Recovery Time Objective during Disaster Recovery with the help of a new approach called Recovery point objective.