Jun Ouyang

About

Backend engineer with DoorDash experience building production-critical systems spanning LLM-based real-time safety moderation (SendBird callbacks + ChatGPT risk scoring with automated actions) and large-scale payments data pipelines (Kafka to CockroachDB with aggregation APIs). Also led cross-team reliability work to standardize SLOs and drove an incident redesign from batch pull to real-time push callbacks to eliminate critical-event latency.

Experience

Education

Key Strengths

• Built and shipped production LLM-based safety moderation with automated enforcement actions
• Improved customer satisfaction scores by 11%+ via real-time message risk evaluation
• End-to-end ownership across BFF/backend, persistence, and actioning workflows
• Cost/performance optimization through experimentation comparing Moderation vs Chat Completions APIs
• Led cross-team SLO consolidation and execution rigor across three engineering teams
• Delivered measurable reliability targets (Tier-0 99.99%; 80% Tier-1 99.9%; Tier-2 99.5%)
• Strong backend/system design for scale (Kafka ingestion, CockroachDB, aggregation jobs, API design)
• Data-modeling decision-making validated via guidance review, platform-team alignment, and experiments
• Incident leadership: reduced crisis-event latency from ~20 minutes to ~2 minutes and redesigned to real-time push callbacks
• Implemented metrics and alerts to prevent recurrence and ensure SLO compliance