Mukul Sabharwal

Experience

• 

  Tata Consultancy Services

  Jun 2011 - Sept 2012

  Business Analyst

  SAS Certified ProgrammerStudy ‘Consumer Shopping Behavior’ across varied FMCG category through markets covering EU, APMEA & LATAM and identify drivers of purchasing decisions and patterns. By understanding competitors shares, suggest Consumer & Product Segmentation, Shelf Layout, Opportunity and Rationalization (if any)]The Nielsen Company being the client, project data comprises ‘Nielsen Homescan Retail Data’ with 10000+ UPCs. Statistical Techniques Utilized: Cluster Analysis, Factor Analysis, Regression etc. via UNIX, SAS and Excel Customized Segmentation:• Member of ‘Customized Segmentation Competency Center’, Nielsen • Responsible for Re-segmenting Market (under study) using 1-2-3 Segmentation model basis facts such as Annual Spending, brand tier etc.• Profiling of segments formed as a result of ‘Cluster Analysis’ both agglomerate and non-agglomerate techniques• Developed statistical models for Product and Consumer Segmentation based on Cluster Analysis • Applied advanced SAS and SAS Macros to deliver customised requests of clients Specialties: SAS Certified ,Base SAS, SAS Macros, SAS Advanced ,SQL, Core Java, UNIX Show less

• 

  KiE Square

  Sept 2012 - Aug 2013

  Business Analyst

  Offshore visit:Study to reduce the forced churn for a telecom client based in Malaysia (KL) Attack TOS 1- way proactively and reduce force churn rates Aid proactive action with relevant insights on TOS 1- way  Prioritize customers who should be called in addition to IVR Defined “Build”, “Test”, ”Score” modeling windowsUnderwent rigorous training of a modeling tool KXEN to study and generate dynamic models in telecom domain

• 

  Dunnhumby

  Sept 2013 - Mar 2022

  Tesco is present in multiple verticals , retail being the key business. Currently I am working in a role that looks at the stores and Channels side of the business as well as the Telecom business of Tesco. Key projects undertaken are:Acquisition ModelingHelped the client develop an understanding of their customers through profiling and propensity science. Mathematics being at the core of machine learning has literally changed everything. So, when we talk about propensity science, what we are usually referring to is a machine learning based approach, that is likely to reduce our cost by reaching out to the right set of customers.With that objective, often, we will need to estimate the right trade-off of having “good” & “enough” data. To arrive at good data, we test our hypothesis (variables/assumptions) with the likelihood of a customer to get acquired along with finding the right value for enough (customers) data.This project has laid a foundation of many other related projects which we will be picking up soon.Sentiment and Sales : Details to be shared soon.Certifications: Since learning is innocuous and knowledge is freely available , but getting certified in a few is necessary and beneficial for our own good. So completed the following:(MOOC's)Machine Learning ( See accomplishments section )Maths behind Machine Learning- Imperial College London ( See accomplishments section )Neural Networks and Deep Learning ( See accomplishments section )Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization ( See accomplishments section )Convolutional Neural Networks (See accomplishments section) Show less Majority of the work during this period was done for a client based in Malaysia. Work spanning across the following domains:Quantitative Research: a. Build attitudinal segmentation to understand what are the important factors in deciding why they chose a particular store over another.b. Analyse what drives a customer to purchase a juice brand over another, also identify the potential opportunity of customers who lapsed.Segmentations:Segmented customer / transactions into meaningful clusters to have a succinct view of the data. Using various machine learning algorithms.Lapsing BehaviourVisited the market for than 3 times to share and work closely with Marketing Director. While working on this project , we tried to delayer the effects of the various levers/channels that impacts the behaviour of customers , such as pricing / Availability / Range Changes in stores etc.Predictive Analysis:Using a customers broad in-store engagement with the retailer, we tried to predict the likelihood of him going online and thus enhancing our understanding of the profile of such customers. Show less As soon as I Joined DH , I was working with the science team here to develop and enhance our understanding of customers missions. i.e. Every time a customer visits a particular store , he / she always has a predetermined budget and an objective in mind.We tried to unearth their in-store behaviour by understanding the kind of missions with which a customer visits a particular store. And size them accordingly. Broadly segmented the missions pattern into 6-8 distinct segments. This work laid the foundation of how we understand missions in the business and soon was deployed in multiple markets. Show less

  • Data Science Manager

    Aug 2018 - Mar 2022

  • Senior Applied Data Scientist : UK

    Dec 2017 - Jul 2018

  • Lead Analyst : Russia

    Apr 2017 - Nov 2017

  • Lead Analyst : Asia

    Jan 2016 - Mar 2017

  • Senior Analyst - Asia

    Apr 2014 - Dec 2015

  • Senior Analyst : Science Team

    Sept 2013 - Apr 2014
• 

  Metyis

  Mar 2022 - now

  • Director

    Jan 2024 - now

  • Principal

    Mar 2022 - Dec 2023
• 

  Metyis

  -

  Director