Giovanni Pacman
Back to the Greenhouse 🌵
AI-Powered Business Intelligence Solutions
AI/ML API BigQuery GCP Unstructured Data

Summary

During my time working with Google Cloud Platform (GCP), I led and contributed to a series of projects focused on building AI-based solutions tailored to specific business needs. By leveraging tools like Google AutoML, Vision AI (Product Search), and Cloud Natural Language API, I helped organizations unlock hidden value from their data, automate complex workflows, and make smarter decisions through enhanced analytics capabilities.

Context & Challenge 🧩

Many companies hold vast amounts of underused unstructured data—images, reviews, product catalogs—due to the complexity of processing it. The challenge is to build scalable, maintainable solutions to:

  • Classify feedback and extract sentiment.
  • Analyze product images to enhance searchability.
  • Make insights accessible to non-technical users.

    • These solutions had to be reliable, cost-effective, and adaptable across industries and data types.

    My Role & Contributions 🧑‍💻

    The work centered around designing and implementing AI solutions built entirely on Google Cloud's native services, ensuring scalability, maintainability, and performance.

    • Google AutoML -> Used to train custom ML models for tasks like ticket classification and churn prediction—enabling fast, code-free experimentation tailored to each client’s needs.
    • Vision AI – Product Search -> Enabled image-based product recognition by finding visually similar items from photos. Optimized for real-world variability (e.g., lighting, backgrounds), it significantly improved search accuracy and user experience, especially in retail.
    • Cloud Natural Language API -> Performed sentiment analysis and entity extraction from unstructured text (e.g., reviews, support tickets), providing real-time insights into customer satisfaction and recurring issues.

    Data Platform Architecture

    These solutions were integrated into production environments with automated pipelines and orchestrated workflows to ensure reliability and ease of use.

    Outcomes & Learnings 🚀

    These projects delivered tangible value by bridging cutting-edge AI capabilities with practical business needs.

  • Automated manual data tasks, cutting time-to-insight and reducing overhead.
  • Turned complex, unstructured data into actionable strategies.
  • Showcased how AI/ML can solve real-world problems at scale.
  • Balanced innovation with performance, cost, and maintainability.

    • They reinforced the power of thoughtful, strategic AI to unlock new opportunities.