Looking for a job

yajna Pale

@yajnapale

Data Scientist at City Logistics who analyzed 214M scan events and built delivery-location validation analysis across 59,897 deliveries.

South Africa

Message

At City Logistics, I analyzed 214M scan events to identify operational patterns and mis-scan rates for Exco decision-making. I also built geospatial delivery-location validation analysis comparing master coordinates with ePOD GPS positions.

Previously, I rebuilt a credit card refund model in Python at Capitec Bank, resulting in approximately R20 million in refunds across 11,000 clients. At Takealot Group (Superbalist.com), I developed churn, customer lifetime value and demand forecasting models, contributing to reduced churn and inventory waste.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

City Logistics

Jul 2026 - Present (4 months)

Analysed 214M scan events and built geospatial validation and route/SLA discovery pipelines to support operational decision-making. Worked across Azure Synapse, SQL Server, Python and PySpark.

Python SQL PySpark Azure Synapse SQL Server Geospatial analysis Data Analysis Machine Learning

Independent Data & Software Consultant

Freelance

Apr 2025 - Jun 2026 (1 year 2 months)

Designed and deployed custom invoicing and billing systems for small businesses, and delivered analytics for a modular container manufacturer. Acted as sole analyst and developer on each engagement.

Python SQL System Design Data Analysis Reporting Documentation User Training

Modelling Specialist / Quantitative Analyst

Capitec Bank

Jul 2021 - Mar 2025 (3 years 8 months)

Rebuilt a manual Excel-based credit card refund model in Python, resulting in approximately R20 million in refunds. Built Power BI dashboards and prototyped migration of credit budgeting models from SQL to Python.

Python SQL Power BI Data Analysis Model Validation Financial Modeling Excel

Data Scientist

Takealot Group (Superbalist.com)

Oct 2019 - Jun 2021 (1 year 8 months)

Developed machine learning models for churn prediction, customer lifetime value estimation and demand forecasting. Reduced customer churn by 12% and improved demand forecasting using ARIMA and Prophet.

Machine Learning Python Churn Prediction Customer Lifetime Value Demand Forecasting ARIMA Prophet

Trade Analyst

Takealot Group (Superbalist.com)

Jul 2018 - Oct 2019 (1 year 3 months)

Rebuilt an offer selection process from Excel to SQL, improving scalability and reducing campaign setup time. Designed and executed flash-sale strategies to move aged stock and increase traffic.

SQL Excel Data Analysis Market Analysis Campaign Management Flash Sales