- Store inventory replenishment model for UK (DS Lead)Developed catboost demand model to predict store sales for UK outlet storesDeveloped pipeline to interface with existing inventory system to establish optimal Weeks of Cover for entire suite of productsLed a team of 2 Associate Data ScientistsStockouts reduced by 10%, Revenue increased by 15%

- Demand forecasting for North America and AsiaEngaged with stakeholders and end-users to determine product specifications and development opportunitiesAdopted and led improvement of time-series models for demand prediction for North America and Asia marketsDeveloped exploratory tools to identify statistical repeat numerical and categorical featuresImplemented hyperparameter tuning in optuna which reduced tuning time from 7 days to 2 hours whilst model MAPE reduced by… Show more - Demand forecasting for North America and AsiaEngaged with stakeholders and end-users to determine product specifications and development opportunitiesAdopted and led improvement of time-series models for demand prediction for North America and Asia marketsDeveloped exploratory tools to identify statistical repeat numerical and categorical featuresImplemented hyperparameter tuning in optuna which reduced tuning time from 7 days to 2 hours whilst model MAPE reduced by >20% which increased planner adoption by 15% Show less

Mentored groups of fellows on the development of machine learning models.

Project #2: Levi Strauss & Company Machine Learning Bootcamp- Designed curriculum (slideshows, notebooks and lab style demonstrations) for bootcamp- Worked with Platform Engineering to identify required access and resources needed for bootcamp participants and developed onboarding documents- Interviewed and assessed participants for bootcamp- Led the instruction for LSA (North America) trackProject #1: Dockers Email Recommendation Project- Developed and optimized… Show more Project #2: Levi Strauss & Company Machine Learning Bootcamp- Designed curriculum (slideshows, notebooks and lab style demonstrations) for bootcamp- Worked with Platform Engineering to identify required access and resources needed for bootcamp participants and developed onboarding documents- Interviewed and assessed participants for bootcamp- Led the instruction for LSA (North America) trackProject #1: Dockers Email Recommendation Project- Developed and optimized email audience selection models for Dockers- Wrote code for deployment of email audience selection model on AWS- Developed a webpage-based tool to optimize the visual content of email campaign creativesTools used: AWS (EC2/ECR/S3/Redshift/Sagemaker/Batch/Lambda/Cloudwatch), Random Forest, Xgboost, Model stacking, Analysis (Partial Dependency Plots, Parallel Coordinates, Correlation Matrices), JIRA / Confluence, SQL, Linux / Bash CLI Show less

Highlights:- Researched and developed machine learning models using agile software development methodology under the mentorship of experienced ML practitioners.- Worked with computer vision and NLP models in the context of fintech and fashion industry.- Participated in weekly AI reading groups to keep abreast of latest developments in the field.- Scrum master / coordinator for group working on multimodal product search models.- Assisted in recruitment of fellows by reviewing… Show more Highlights:- Researched and developed machine learning models using agile software development methodology under the mentorship of experienced ML practitioners.- Worked with computer vision and NLP models in the context of fintech and fashion industry.- Participated in weekly AI reading groups to keep abreast of latest developments in the field.- Scrum master / coordinator for group working on multimodal product search models.- Assisted in recruitment of fellows by reviewing code challenges and video interviews.Tools used:Pytorch, Fastai, GIMP, Scrapy/Splash, Docker, NLP (BERT/ULMFit), Google Hire Show less

Control SystemsAutomationRoboticsResearch Methods