Talha Qayyum
Data Scientist
Islamabad, Islamabad Capital Territory, Pakistan
5+ Years Exp
Summary
Talha Qayyum is a brilliant Data Scientist with more than 4 years of experience combining data with Machine Learning to inform strategy and business decisions. Talha currently leads a data science team working in collaboration with cross-functional teams to develop interactive, real-time dashboards for customer segmentation, risk analytics, and KPI tracking with Power BI. He developed a credit scoring model based on customer demographics using ML and statistical techniques and developed a deep learning model for customer sentiment analysis. Talha is responsible for enabling ML services including bug fixing and troubleshooting. His technical skills include Python, SQL, PostgreSQL, matplotlib, Spacy, PyTorch, Numpy, TensorFlow, Jupyter, Anaconda, AWS, and OpenCV. Talha is passionate about bringing order to vast, messy data sets to reveal powerful insights through deep learning, NLP, ML, and AI techniques.
Technical Skills
Detailed View
Python
Pandas
NumPy
Spacy
PostgreSQL
SQL
Anaconda
Jupyter
AWS
Git
Clustering
Classification
Regression
Naïve Bayes
Decision Modeling
Work Experience
Data Scientist
Credit Per Pvt Ltd
Full Time | 18/03/2021 - Present
Islamabad, Islamabad Capital Territory, Pakistan
- Leading Data Science department, task management in Agile platforms.
- Developed Credit Scoring Model based on customer demographics to calculate customer credit worthiness leveraging various statistical and machine learning techniques.
- Collaborated with BI, Marketing team and designed interactive real time dashboards on PowerBI with SQL, PostgreSQL database integration to analyze customer segregation, Risk Analytics and Company KPI`s.
- Designed and developed RaaS (Risk at Scale) Framework to identify risks involves with each customer.
- Exposure to AWS, managing production environment and database storage.
- Involved closely with external stakeholders to understand the business problem, communicating their requirements to development team and demonstrating progress throughout the projects through reports and dashboards.
- Responsible for monitoring and maintaining machine learning services for any bug fix and troubleshooting.
Skills: Python, Flask, Amazon EC2
- This deep learning model helped the company to evaluate customer credit worthiness based on various demographic and financial behavior.
- Built the model which has highest accuracy in classifying the feedback as approved, rejected or escalated(under-review).
Skills: Python, Pytorch, tensor flow, googleOCR
- Implemented document verification module using Google Vision OCR for Credit Fix Mobile App.
- The module is able to extract fields from utility bills and employment letters for specific verification purposes.
Lecturer
Isra University
Full Time | 05/01/2013 - 25/03/2021
Pakistan
IEEE Conference Paper Title:
- A distributed approach for solving the position inverse kinematics of mobile parallel manipulators.
Lecturer
Isra University
Full Time | 17/06/2016 - 21/03/2021
Islamabad, Islamabad Capital Territory, Pakistan
- Teaching engineering courses i.e., Machine Learning, Artificial intelligence, and electronic devices.
- Developed proposals for major research projects which will make a significant impact, and lead to an increase in knowledge or understanding or the development of new explanations, insights, concepts or processes.
Machine Learning Consultant
Metis International Pvt Ltd
Full Time | 15/06/2017 - 15/02/2021
Islamabad, Islamabad Capital Territory, Pakistan
- Developed time-series forecasting model for fintech client using Cross validation technique and using Random forest classifier then evaluated algorithms to improve performance, quality, data management and accuracy.
- Designed and implemented new machine learning based approaches for predictive modeling, developed high precision classifiers and tools leveraging machine learning, regression and rule-based based models using Amazon Sage maker Services (AWS Amazon).
- Improved Data Quality Management techniques to improve data collection from customer form filled data.
- Developed SMS parser to parse mobile data (messages) to compile financial statement using Bank messages.
Skills: Python, Spacy, googleOCR, Amazon EC2
- Developed a machine learning based model which initially trained with various fraudulent customer anomalies later used to predict whether the customer fraudulent behavior based on his previous credit history and enquires.
Skills: BERT, Python, NLTk
- This deep learning model helped the company to understand better about their customer feedbacks so that they can concentrate on those issues’ customers are facing.
- Build the model which has highest accuracy in classifying the feedback as positive, Negative and neutral.
Education
Master of Science
National University of Science and Technology
Major in Robotics and Intelligent Machine engineering
Bachelor of Science
International Islamic University
Major in Electronics Engineering
Certifications
Machine Learning with Python
Coursera
Deep Neural Network with PyTorch
Coursera
Natural Language Processing Specialization
Coursera
Machine Learning with TensorFlow
