Contact
Contact

Event Clock Machine Learning

AI/ML
Back to Projects

Developed an adaptable machine learning model to identify and significantly reduce false positives within event tracking systems. This project directly enhanced system performance, reduced operational costs, and improved overall customer satisfaction through more accurate event data.

Problem & Solution

The Challenge

The presence of false positives in event tracking data led to considerable inefficiencies, misallocation of resources, and posed a risk to customer trust due to inacurrate logging and alerts.

My Solution

The project team consisting of my co-intern and myself engineered an end-to-end ML pipeline that extracted batch data from BigQuery, meticulously organized it into training-ready datasets, and generated predictive models using Google Cloud AutoML. Model performance was continuously fine-tuned by adjusting confidence thresholds to align percisely with critical business objectives.

Project Info

Duration: 3 Months

Role: Software Engineer

Technologies Used
BigQuery
AutoML
VertexAI
Tableau
Jupyter Notebook

Key Features

  • Machine Learning Pipeline Development

  • AutoML Model Training & Prediction

  • Tableau Dashboard for Insights Visualization

  • BigQuery Data Extraction & Preparation

  • Custom Confidence Threshold Tuning

  • False Positive Reduction System

What I Learned

  • Balancing multitasking for project deadlines

  • Understanding domain-specific modeling challenges

  • Building scalable ML solutions on GCP

  • Applying data visualization for insights