Product Managed:
• Lightweight user feedback experience products for ads across FB, IG, etc. to capture right type of feedback. Launch increased user feedback exponentially (5-10x).
• Reactive / Secondary review system that acts as a second line of defense to detect ads that violate FB's Ad policies. Launched a specialized review platform with a custom machine learning service that supported heuristic based triggers, candidate selection, ranking, prioritization and ML assisted human review. Optimized for recall, efficacy and maximum RoI, it increased contribution to overall prevalence reduction with a fraction of overall human review capacity and also increased fully automated enforcement. Launched novel concepts like Content Summarization, ML Model Reward System (to reward models that performed better), Automating video review, ML assisted routing, etc.
• Platform to down rank low quality and non compliant ads through ML based re-review and prioritization to maximize system & labelling efficiency to reduce bad user experiences. This platform resulted in exponential improvements to recall, precision, training data quality and reduced feature sparseness
• Plug and Play user sentiment measurement platform to identify new opportunities for delivery improvements or enforcement and re-evaluating existing enforcement strategies and org priorities
