Software Engineer
Meta
Promoted from E3 to E4 · Privacy and production ML workflows
June 2022 - Nov 2025 · Remote
- Built and operated latency-sensitive real-time ML classification pipelines for privacy workflows, supporting 10K+ QPS production workloads with staged validation, rollout testing, and zero-downtime transitions.
- Evaluated real-time serving data availability against model input requirements, identifying coverage, quality, and edge-case gaps while onboarding required production data sources.
- Improved observability and on-call reliability by expanding metrics, alerts, and dashboards, reducing incident detection time to under 5 minutes and simplifying debugging workflows.
- Built a Python-based automation workflow that reduced manual review time by 50%+ while improving consistency across production review tasks.
- Built ML API integration layers enabling offline workflows at tens-of-millions scale to safely access rate-limited real-time serving systems.
- Reprioritized and delivered a multimodal ML backend integration with two partner teams, shipping production workflow changes alongside the scheduled roadmap.
- Owned migration of ML workflows to a new serving platform, improving reliability and rollout speed with step-by-step serving documentation adopted by partner teams.