Currently I lead the strategy and execution behind the One Order Forecast platform, which is an enterprise level forecasting engine used to plan over 10M + orders each month in 15 countries and 500 cities for the JET group using a Human-in-the-loop process for scenario planning and explainable predictions. Built the Machine Learning team from scratch. Order Volume and Courier Demand Prediction using Machine learning algorithms for JustEat Takeaway Scoober(own delivery) operations for 200+ cities in Europe. Tech stack: Python (numpy, pandas, xgboost, pytorch, scikit-learn), Docker, Airflow, AWS - (RedShift, Sagemaker, S3, Lambda, API Gateway), StreamLit, Weights&Biases, DVC, SQL, R (ggplot2, dplyr) Collaboration tools: JIRA, Confluence, Miro, Git, DVC, Weights&Biases Responsibilities: involved hands-on from ideation to production of analytics and ML solutions in forecasting/routing/ETA prediction, lead hiring from job description to final offer, building analytics and AI capabilities within Scoober through collaboration with counterparts in London, Amsterdam, and Winnipeg on key MLOps topics for Takeaway group.

Omio | Jan 2019 - Dec 2019

Data Scientist - Revenue and Monetisation

Built data driven approach to forging sales partnerships in fast growing revenue domains - Central and Eastern Europe(fragmented bus market) and Ferries. Ran and analysed AB tests on booking fees to improve company bottom line. Built a data-driven way of measuring team’s sales performance

PwC | Aug 2016 - Oct 2018

Management Consultant

Over 2 years at PwC, I was involved mostly in projects related to improving business operations leveraging data science. Most of the engagements were quantitatively rigourous which required quick turnaround times. I worked with clients like Standard Insurance where we helped transition them from their 40-sheet Excel based reporting to PowerBI based dashboards powered by automated R scripts in the backend to get a quick high level business overview as well as deep-dive into revenue and profitability metrics. Helped doctors at Health care behemoths like MD Anderson and Mayo Clinic to understand impact of cancer drug interventions using survival analysis through interactive RShiny dashboards hosted on web servers. Built and productised credit-risk models for Hapaolim bank In Israel in order for them to better understand which of their customers were most likely to default. Enabled jetStar, an Australian carrier, better understand the impact of a service fee levy by Singapore’s Changi airport on its business and the industry using scenario modelling. Tech stack: R (dplyr, caret, tidy verse), PowerBI, Docker Techniques used : Time-series modeling, ARIMA and ARIMAX modeling, Decision Trees, Random Forest, Gradient boosting, Survival Analysis and Ensemble Methods.

Fractal Analytics | May 2013 - Dec 2013

Business Analyst

In my first job, I analysed the performance of P&G’s Pampers brand in the APAC region. At the time, Pampers was a billion dollar brand in terms of revenue for P&G. Pampers was gradually losing share to Japanese Mamy Poko Pants brand and our task was to analyse what was the competitor’s edge. MPP was not a top tier brand like Pampers and was able to launch lower tier products in segments where Pampers was not present/did not want to enter. Decided to quit the job to learn programming, statistics and econometrics(modelling).

My Education

Take a look at my formal education

Indira Gandhi Institute of Development Research

St. Stephen's College

Delhi Public School - Mathura Road

Create Your Data Portfolio for Free