- Developed SQL and Python pipelines that supported analytics for support operations, resulting in a 50% reduction in manual data handling

- Collaborated cross-functionally with Product and Customer Success teams to design A/B tests; results guided product feature prioritization

- Created in-depth customer engagement dashboards with PowerBI, DAX and MySQL to identify support process bottlenecks, providing a real-time view of KPIs and aiding optimisation strategies which led to a 15% increase in productivity

- Refactored and productionised the data pipeline powering ML-driven features such as Lead Scoring and Deal Close Likelihood Prediction in Freshsales CRM, enhancing both model accuracy and pipeline reliability.

- Maintained the schema layer in a multi-tenant distributed data processing engine which handled 800k messages per minute

- Collaborated cross-functionally with finance, customer, and product teams to deliver insights and platform improvements

- Trained and mentored 3 analysts, fostering a culture of collaborative analytics and best practices in storytelling with data

Work Integrated Learning

Data Science Engineer | RMIT Marketing Ad Analytics Platform | Aug 2024 - Oct 2024

RMIT University

• - Led a data-driven ad analytics project for RMIT's marketing team, improving digital campaign click-through rates by 10%. This involved analyzing campaign performance (Google Ads), developing web-based dashboards to visualize key metrics, and providing recommendations for A/B testing ad creatives and targeting strategies.
• - Designed and architected an efficient, event-driven, Python-based platform, resulting in 50% reduction in AWS costs, specifically by switching to RDS from DynamoDB
• - Reduced research cycle time by 90% by automating ad collection, processing, and insight extraction through an ETL pipeline.
• - Owned auth and data visualisation modules and delivered a secure platform with Role-Based Access control
• - Optimised API response times from 25 seconds to under 4 seconds, resulting in a 5x performance improvement in the competitor analytics module
• - Achieved a 99% retrieval precision rate, ensuring that most top-N results were relevant to user queries, by utilising PGVector in Postgres and text embeddings instead of raw text-based search
• - Streamlined CI/CD processes, cutting deployment errors by 80% using GitHub Actions

Part Time

Customer Service | Nov 2023 - Present

Woolworths Supermarkets

Human Resources Manager | September 2019 - August 2020

TEDx Chennai