Hello!
I’m a Data Scientist with over four years of experience specializing in data analytics, machine learning, deep learning and MLOps. I’ve worked in different places from consulting firms to dynamic startups, where I’ve gotten better at what I do and helped with various projects.
Some of cool things I’ve done include predicting if customers will leave, testing which discount vouchers work best, figuring out credit scores, analyzing sentiments, OCR (extracting text from ID card), monitoring important numbers, designing datawarehouse with Data Vault model.
I’m good at using tools like python, bigquery, tableau. Also know my way around airflow, Rest API, and familiar with GCP.
Feel free to reach out if you have any position for which you may consider me to be a suitable candidate
Look at my CV
Portfolio
- Churn Prediction
- Credit Scoring
- Sentiment Analysis Pipeline
- Restaurant Recommendation
- OCR for Student ID Card
- Kubeflow House Prediction Pipeline
Churn Prediction
This project focuses on predicting customer churn utilizing machine learning using their past behavior. Additionally, the project includes the development of a RESTful API.
category:
Credit Scoring
Developed a credit scoring model for Company ‘X’ to assess customer credit worthiness based on data from past borrowers. The aim is to help the company meet their target loan approval rate and default rate.
category:
Sentiment Analysis (App Store Scraping)
This project employs Airflow and KubernetesPodOperator to orchestrate a seamless pipeline for scraping user reviews from app store, conducting sentiment analysis, also providing with example dashboard.
category:
Restaurant Recommendation
This project helps people in Nashville find great restaurants easily. We use a hybrid method to make personalized recommendations for users, collaborative filtering and content-based filtering. I also handle cold start.
category:
Text Extraction From Student Card (OCR)
Developed an OCR system for Dawnvale Academy of Medicine student cards. It detects card corners, adjusts orientation if needed, and extracts name, member ID, address, and phone number. Utilizes similarity measures against a district master dataset for accuracy.
Category:
Kubeflow Pipeline - House Price Prediction
This project solely focuses on leveraging Kubeflow Pipelines for house price prediction, from preparing data, training model to saving it to google cloud storage. Also includes deploying on kubernetes without getting into the details of methodology of building models.
category: