Back to Home

ShortMax - Personalized AI Recommendation System

The Problem

As the content catalog scaled, generic discovery made it harder for users to find relevant shows, reducing engagement and retention. Meanwhile, there was no clear metrics framework to measure content performance — creating a growing need for data-driven, personalized recommendations.

Solution and Contribution

I led the development and demonstration of an AI-driven recommendation system to personalize content discovery. Established key business metrics system across CTR, watch time, retention and conversion to evaluate app performance. Bridged Product and Engineering to turn user behaviors into data-driven decisions, driving stronger engagement and subscription revenue.

Timeline

March 2024 - May 2024

My role

Solutions Engineer; Supporting Developer

ShortMax app screenshot showing personalized content recommendations

Customer Impact Result

+35% CTR
+27% Watch Time
~19% MRR Growth within 3 Months

ShortMax viewer engagement metrics chart showing CTR, Watch Time, and Dwell Time trends

User Research

Using AI algorithms to identify users' wants and needs

Based on user behavior data, customer feedback and cross-functional discussions with Product and Engineering, we identified that users needed to:

  • Discover shows that better match their individual interests
  • Spend less time searching and skipping irrelevant content
  • Develop and reinforce viewing preferences through personalized AI feeds
  • Discover new content aligned with their evolving tastes

Solution Discovery

How do we identify the right metrics to capture customer preferences?

To build a meaningful evaluation framework, I worked across Product, Engineering, and business stakeholders to identify which signals best reflected user preferences through:

  • Product discussions and behavioral insights
  • Backend data and signal analysis
  • A/B testing and metric validation
  • Stakeholder alignment and demo

Implementation and Iterations

The process: Build → Measure → Align → Iterate → Scale

Throughout implementation, I communicated closely with Product, multiple Engineering teams and stakeholders to

  • monitor user data
  • refine the metrics framework
  • support technical implementation
  • demo iterations

to ensure the solution stayed aligned with business goals and drove revenue growth.

A/B testing dashboard and code implementation for the recommendation system

A/B Testing and Code Implementation

Competitor analysis presentation comparing market landscape and AI solution

Competitor Analysis Presentation

Tech and Tools

Python · React · SQL · AWS · REST APIs · Git · Canva