Chetan Halai

Actively looking for an internship or Junior Data Science/Analyst role 

Have done well versed in owning end-to-end machine learning projects, experienced in analysing large, unstructured data sets (with 50million + rows) building classification/regression models, large scale Time Series projects and able to deploy models onto Streamlit and AWS.

Projects

• New York Taxi  - 
• Built a complete ML service – to predict rider demand and driver availability to help encounter driver partners  - Merged 1 year of taxi tip data (50 million + rows of data) . Converted the raw data to tabular for Supervised ML, applied Regression, XGboost and Light GBM. Finally, I operationalised the model, using Hopsworks feature store and deployed it on to Streamlit
• Techstack: Scikit- Learn, Hopsworks, Streamlit, Xgboost, LightGBM, Plotly, Pandas, Numpy
• Ola Demand Forcast 
• Project required me predict demand for rides in a certain region for a particular latitude and Longitude for a requested future time window. Imported (8million rows of data), Cleaned/Pre-processed, refactored features for model ingestion, conducted Geospatial Feature Engineering - employing "K-Means Clustering" to partition India into distinct regions, to create unique Location ID’s. Deployed Random Forest regressors and XGboost and then built a prediction pipeline ready to deploy .
• Techstack: K-Means, Pandas, Matplotlib, Random Forest, XGboost, ellipsoidal model.
• Loan Default Prediction with Machine Learning
• Cleaned and wrangled bank loan data to extract insight for decision making. Built machine learning ensemble models to predict customers who are worthy to be given a loan and those that are not. Helped identify important variables that influence the eligibility of a customer for a loan. (voting classifier- need to add)               Techstack: AWS Sagemaker, Boto3, S3 Bucket, Pandas, Numpy, Seaborn, Matplotlib, data Visualization
• Multi-touch-Attribution Model and Marketing Spend Optimisation
• Built multiple attribution models to identify the ROI of various marketing efforts and their impact on conversions or sales. This enabled management business to make decisions based on the millions of converting click paths by isolating the impact of every touchpoint.
• Completed the DataCamp Job Ready Profesional Data Scientist Certification 
• Extensive 9 month course finished with timed assessment marked manually by inhouse data science experts, . I was also required to do a presentation tailored to a nontechnical audience, who were interested in why the work was done and what the outcome was. 

I have recently decided to do a career switch into Data Science - I have recently taken time out to enhance and establish my skillset with a view to securing my first Internship or Junior role as a data scientist. 

I have completed the following projects to equip myself with skills to be a data scientist: 

Actively looking for an internship or Junior Data Science/Analyst role 

Have done well versed in owning end-to-end machine learning projects, experienced in analysing large, unstructured data sets (with 50million + rows) building classification/regression models, large scale Time Series projects and able to deploy models onto Streamlit and AWS.

Projects

• New York Taxi  - 
• Built a complete ML service – to predict rider demand and driver availability to help encounter driver partners  - Merged 1 year of taxi tip data (50 million + rows of data) . Converted the raw data to tabular for Supervised ML, applied Regression, XGboost and Light GBM. Finally, I operationalised the model, using Hopsworks feature store and deployed it on to Streamlit
• Techstack: Scikit- Learn, Hopsworks, Streamlit, Xgboost, LightGBM, Plotly, Pandas, Numpy
• Ola Demand Forcast 
• Project required me predict demand for rides in a certain region for a particular latitude and Longitude for a requested future time window. Imported (8million rows of data), Cleaned/Pre-processed, refactored features for model ingestion, conducted Geospatial Feature Engineering - employing "K-Means Clustering" to partition India into distinct regions, to create unique Location ID’s. Deployed Random Forest regressors and XGboost and then built a prediction pipeline ready to deploy .
• Techstack: K-Means, Pandas, Matplotlib, Random Forest, XGboost, ellipsoidal model.
• Loan Default Prediction with Machine Learning
• Cleaned and wrangled bank loan data to extract insight for decision making. Built machine learning ensemble models to predict customers who are worthy to be given a loan and those that are not. Helped identify important variables that influence the eligibility of a customer for a loan. (voting classifier- need to add)               Techstack: AWS Sagemaker, Boto3, S3 Bucket, Pandas, Numpy, Seaborn, Matplotlib, data Visualization
• Multi-touch-Attribution Model and Marketing Spend Optimisation
• Built multiple attribution models to identify the ROI of various marketing efforts and their impact on conversions or sales. This enabled management business to make decisions based on the millions of converting click paths by isolating the impact of every touchpoint.
• Completed the DataCamp Job Ready Profesional Data Scientist Certification 
• Extensive 9 month course finished with timed assessment marked manually by inhouse data science experts, . I was also required to do a presentation tailored to a nontechnical audience, who were interested in why the work was done and what the outcome was. 

Prior to this I have worked predominantly as a recruiter

GMI UK – Global Head of Talent Acquisition                                                                                                                     

May 2018 – August 2019 

Achievements: 

• Reshaped GMI’s executive hiring strategy, which led to successfully securing candidates from a direct competitor for roles such as UK COO, Head of Operation (London), Country Head (Singapore, Malaysia and Minks), Head of Sales(London).
• Successfully overlooked the hiring and sourced 43 candidates across Technology, Client Services, Sales and Operation in the UK, Belarus (Minsk), Shanghai, Hong Kong, Malaysia and Singapore.
• Provided structured, regular reporting on candidate pipeline, sourcing effectiveness, time to hire etc., to management to determine challenges and remove barriers to success.
• Developed a talent pool framework that helps the business map out its workforce requirements and work backwards to an efficient, effective recruitment strategy.
• Created an internal referral program to reduce and eventually eliminate all agency costs 

• I have 3 Alevels 
• Completed the DataCamp Job Ready Profesional Data Scientist Certification