Heena Shahanaz Shaik

Innovative Machine Learning Engineer and Natural Language Processing Expert with around 2 years of experience in machine learning model design, development, testing, and deployment. Good experienced in delivering end-to-end machine learning and data science projects with the right choice of technologies, and algorithms to solve real-world use cases in the retail industry.

MACHINE LEARNING ENGINEER AND NLP EXPERT

• Classified the invoice documents into multiple categories based on the document type.

Roles and Responsibility:

• Collected multiple invoices related to multiple products.
• Extracted data from invoices by using regex patterns.
• Used stemming, lemmatization, stop-word removal and Garbage values removal for pre-processing.
• Used count vectorization,TF-IDF to generate the sparse matrix for the analysis.
• Built a model using Random Forest Classifier and hyper parameter tuning is done by using GridSearchCV to get the best parameters for the Algorithm.
• Test the model for a few weeks and after validating the accuracy for those 30 days, pushed the model to the production.
• After getting new data, combined with old documents and keras, tokenizer and word2vec is used for vectorization of text corpus.
• Build a Deep Neural Network model to classify the documents.
• Evaluated the model using a confusion matrix and pushed the model to the production after rigorous testing of the model for one month.

• Detected emotions of customers through analysis of their reviews by adding Sarcastic reviews also.

Roles and Responsibility:

• Used stemming, lemmatization,stop-word removal and Garbage values removal for pre-processing.
• Created a word cloud for visualization and analyzed the sentiments from keywords being used in the review.
• Used count vectorization,TF-IDF to generate the sparse matrix for the analysis.
• Built a model using Naive Bayes.
• Evaluated the model using a confusion matrix and accuracy.
• Test the model for 30days and after validating the accuracy for those 30 days, push the model to production.
• After getting new data, combined with old data and keras tokenizer is used for vectorization of text corpus.
• Build LSTM Classification model to predict the review.
• Evaluated the model using a confusion matrix and pushed the model to the production after rigorous testing of the model for one month.

• Identifying the Right Channel that drives more Sales in the Retail Stores.

Roles and Responsibility:

• Implemented Design Thinking Approach, Prioritization matrix, and data validation Techniques.
• Exploratory Data Analysis is implemented on the base data set to create Analytical Data Set.
• Checked for Linear Relation, Autocorrelation, Multicollinearity, Heteroskedasticity, Homoscedasticity, and Normally distributed errors for satisfying Linear Regression.
• Ordinary Least Squares method is used to predict the coefficients.
• Implemented Dimension Reduction Technique (Principal Component Analysis) on the base analytical data to improve the accuracy by adjusting for the bias and variance trade-off.
• Identified the top 5 drives and bottom 5 drivers and Computed R2, Adj R2, RMSE, MAE, and MSE and shared the results with relevant stakeholders.

• Certification in Coursera on “Introduction to Web Development” with 93.8% in 2020.
• Coursera Certification in “Programming for Everybody (Getting Started with Python)” with 99.08% in 2020.
• Coursera Certification in “Crash Code on Python” with 89.50% in 2020.
• HackerRank Certificate in “Python (Basic)” on 8th July,2020.