T Dao Nguyen

Problem: Production application was running on an older stack (.NET/AngularJS), making it slower to extend and harder to maintain as requirements grew.
Approach: Migrated the application to a modern stack — .NET Core/ReactJS with server-side rendering, static generation, client/server side validation, while preserving existing functionality to minimize downtime during the transition.
Result: Reduced time to ship new features by 97.9% and a 48x speedup, eliminated a class of recurring bugs tied to the old framework, and optimized website SEO resulting in increased organic traffic/improved rankings.

Problem: Systems needed secure access control for both internal and external users, without a prior solution.
Approach: Designed and implemented auth flows using OAuth/OIDC protocols, supporting internal tooling and external partner access with token-based authentication.
Result: Delivered a secure, standards-based auth layer used across internal and external user bases, reducing security risk / manual access-management overhead.

Problem: The same customer frequently existed as multiple records across systems, due to inconsistent entry points and formatting — fragmenting engagement history and skewing metrics.
Approach: Wrote Python-based normalization and matching logic to identify, reconcile, and merge duplicate customer records across connected systems, with validation safeguards to prevent incorrect merges.
Result: ~20,000 duplicate records resolved to date, with the normalization pipeline still active and running continuously.

Problem: Customer data lived in silos across ad platforms, CRM, and organic channels, with no reliable single view of who a customer actually was. Approach: Architected, built, and now own a Customer 360 inside Zoho CRM, consolidating multi-channel data into a single source of truth, with the system designed to scale into a full customer data platform (CDP) over time. Result: A unified customer view now feeds reporting across departments, replacing what used to be fragmented, inconsistent data spread across disconnected systems.

Problem: Teams relied on manually compiling recurring digests from reputable industry sources, research papers, etc. which was repetitive and easy to deprioritize under other work. Approach: Built a scheduled job using Claude to automatically generate marketing and compliance digests on a daily basis, pulling from data sources that are vetted and finalizing digest after a review step to confirm output accuracy before distribution.
Result: Digests now run unattended on schedule, reducing manual reporting overhead and ensuring consistent, on-time delivery ..