Education
- 2023 - 2024
- 2012 - 2016
Experience
-
2020 - 2022
Philips
Data Science Manager
•Spearheaded the development and democratization of Scalable Interpretable Repeatable Auditable Analytics (SIRAA) engine
capable of Market Mix Modelling, short-term forecasting, Promo Effectiveness Evaluation and automatic reports generation saving
1.25 million EUR in productivity costs
•Implemented Bayesian Marketing Mix models and integrated with SIRAA for scalability using Stan, Python and R reducing
attribution error by 20%
•Created a Marketing Simulator engine reporting ROI and media budget allocation recommendations to campaign managers which
led to improved sales of personal care segment by 6%
•Automated data pipelines in collaboration with data engineers capable of extracting, cleaning, re-conciliating, modelling dataset
generation integrated with SIRAA. The data sources include sellout from online, offline retailers, media data from FBBM, Google
Search, SA360, DV360, Instagram Ads, brand Influencers, TV, Amazon Sponsored Ads and Display improving productivity by 50%
•Democratized the process to publish the results to Qliksense dashboards collaborating with business intelligence engineers,
improving the consumption of insights globally by 30% -
2016 - 2020
Mu Sigma
Apprentice Leader | Decision Scientist
•Supervised a team of five decision scientists to evaluate AutoML products and recommend the best fit to the client’s team based on
their infrastructure and projects, reducing the time for model development and deployment by 50%
•Led a team of two analysts to develop a predictive solution encompassing data pipelines to generate payment plans for an
Architectural Engineering firm to manage working capital efficiently and reducing 80% manual effort
•Identified the bottlenecks in the Pilbara port operations of an Australian mining firm and prototyped a predictive solution to
estimate the vessel fill percentage increasing throughput by 6 MTPA
•Scoped future projects by analyzing the data of ore handling plant and supply chain identifying bottlenecks costing 12 MTPA
•Built a Demand Sensing prototype along with an R-Shiny dashboard for an AC manufacturer firm to improve forecast accuracy from
65% to 80% ingesting POS data
•Improved the demand planning process of a Medical Device manufacturer with a portfolio of $102B by building a demand
forecasting engine for global supply and operational planning team and minimizing back orders saving $15MM
•Reduced the refresh time of the forecasting engine by 70% utilizing parallel processing on AWS-EC2 instance and facilitated the
adoption of forecasts by having monthly meetings with demand planners from different regions, collected feedback data further
creating a roadmap of future improvements
•Mentored new inductees on machine learning, first principles thinking, data science and reviewed their mock projects