- Write SQL and do the visualization to understand the user's behaviors and interests.

- Define a target metric to measure user engagement.

- Build a Machine Learning model to classify engaged and unengaged users.

- Setup and analyse A/B Testing to validate the model.

- Design, data modeling,execution and statistical analysis.

- Work directly with business owners and IT to plan, execute, and analyze all A/B & multivariate tests.

-Develop and document testing processes and policies to further increase the quality and rigor.

-Provide analysis of each experiment to understand the impact on marketing campaigns or product roadmaps.

- Analyse and collect unstructured data.

- Database design and data modeling

- Implement ETL and automate the integration process.

- Be responsible for the availability and the reliability of data needs of other departments

- Create and improve tools for data science and data analytics team

- Analyse the data and create new features

- Transform text data into Term Frequency - Inverse Document Frequency, select the best feature with f_classif and fit the transformed data to Bayesian algorithm

- Transform text data into Word Embedding, select the best feature with f_classif and fit the transformed data to Bayesian algorithm

- Word Embedding + Bayes improves 8?curacy from 86% (baseline) to 94% Meanwhile TF-IDF + Bayes improve 5?curacy from 86% to 91%

- Github:https://github.com/diem-ai/text-classification

I built Native Bayes classi er on Term Frequency–Inverse Document Frequency and Word Embedding models andused one of Selection Best Feature techniques to choose only features that contribute to the performance ofthe prediction. The accuracy score is 94%. It is 8% better than the baseline (86%).

I built an AI Assistant that helps customers to place their pizza orders. The machine learning model iscreated by training Rasa NLU from intents , entities, stories, responses, rules and actions. The bot canunderstand the similar messages by extracting the entities and handle out of scope messages.

I wanted to analyze customers interaction throughout the customer journey so that Im enabled to learnsomething about customers and thereby improve marketing opportunities and purchase rates.I used Principal Component Analysis and Agglomerative Clustering algorithm implemented by Scikit-learnto describe the variation in the di erent types of customers that a wholesale distributor interacts with andto predict dynamically segmentation so that I can design campaigns, creatives and schedule for productrelease

I developed a system that is able to learn the movie preferences of users through their ratings and moviecontent. I built Collaborative ltering and Content Based Filtering models by using ratings andcharacteristics of the users and items.

The goal is to explore logistic regression and feature engineering with existing sklearn functions and unpackthe topics in customers review. with Latent Dirichlet Allocation. I used product review data fromAmazon.com to predict whether the sentiments about a product (from its reviews) are positive or negativeor neutral.