Summary
Overview
Work History
Education
Skills
Timeline
Generic

Priyanka Vijay Lahoti

BANGALORE

Summary

Senior Data Scientist with 4+ years of experience in building machine learning models on large scale data to drive business decisions.

Overview

4
4
years of professional experience

Work History

Data Scientist (G4)

PhonePe Pvt Ltd
Bangalore
11.2022 - Current
  • Developed and deployed document detection and masking models (ID Card Detection, Aadhaar Masking) to replace external vendor IDFY, eliminating per-API-hit charges and streamlining production workflows.
  • Built OCR models for PAN, Driving Licence, Voter, Passport, and Aadhaar cards, enhancing KYC processes with improved accuracy and reduced latency through multiprocessing, saving costs and optimizing compliance.
  • Engineered a name-matching algorithm for KYC across multiple business areas, using regex and dynamic programming to increase accuracy and speed. Created a Streamlit app for stakeholders to set match thresholds.
  • Led Video KYC liveness detection development using tflite, enabling real-time compliance checks in 2-3 seconds.
  • Developing sanctions screening and proof of business verification algorithms to automate and secure onboarding, reducing manual errors and vendor costs.
  • Collaborated with cross-functional teams (product, engineering, ml-platform) to deliver high-impact models and streamline deployment on PhonePe’s ml-platform.

Data Scientist (L3)

Gap Inc
San Francisco
04.2021 - 11.2022
  • Demand Forecasting: Improved the accuracy of demand forecasting models by 3-5% for online and retail channels, addressing model bias with a custom Huber loss function.
  • Mentorship: Mentored an intern on building a demand forecasting model using encoder-decoder LSTM architecture.
  • Product Similarity POC: Developed a proof of concept to identify similar products using only images, leveraging K-means clustering and CNNs in TensorFlow.
  • Model Optimization: Reduced SAS model testing time by 50% through Linux-based automation.
  • Performance Measurement Pipeline: Designed a Spark SQL-based pipeline to compare model performance weekly on millions of records, enabling faster, data-driven decisions for brand partners.

Data Scientist (Intern)

Ralph lauren
New York
09.2020 - 12.2020
  • Return Propensity Model: Developed a machine learning model in AWS using Python to predict item return propensity, leveraging CatBoost regression and logistic regression for enhanced accuracy.
  • Model Explainability & Calibration: Utilized SHAP values for model interpretability and calibrated binary classification probabilities to reflect true return likelihood, improving decision-making accuracy.

Data Scientist (Intern)

Assurant
New York
06.2020 - 11.2020
  • Image Classification & Object Detection: Developed a crack and scratch detection model using UNet in Keras on Azure Databricks, achieving 90% precision in identifying screen defects.
  • Call Volume Forecasting: Collaborated with Assurant's Global Automotive team to improve call volume forecasting across 11 contact types, reducing forecasting error by 70% (from 10% to 3%) using ARIMA, ETS, and LSTM models in Azure Databricks.

Education

Master of Science - Data Science

Columbia University in The City of New York
New York, USA
12-2020

Master of Science - Mathematics & Scientific Computing (Gold Medalist)

NIT Warangal
Warangal, India
05-2018

Skills

  • Machine Learning
  • Deep learning
  • Python, SQL, Spark expertise
  • Linear algebra
  • Statistical analysis
  • Probabilistic models
  • Problem-solving abilities
  • Teamwork and communication

Timeline

Data Scientist (G4)

PhonePe Pvt Ltd
11.2022 - Current

Data Scientist (L3)

Gap Inc
04.2021 - 11.2022

Data Scientist (Intern)

Ralph lauren
09.2020 - 12.2020

Data Scientist (Intern)

Assurant
06.2020 - 11.2020

Master of Science - Data Science

Columbia University in The City of New York

Master of Science - Mathematics & Scientific Computing (Gold Medalist)

NIT Warangal
Priyanka Vijay Lahoti