가영 김

I've built and operated production backend services for Muzlive's K-pop artist platform using Python, Django, PostgreSQL, and AWS. I own work from investigation and design through QA, deployment, monitoring, and documentation.

I completed a profile-to-account architecture migration end to end and reduced core analytics API latency from seconds to milliseconds by replacing its underlying algorithm. I also built Gemini image generation APIs with failure fallback so rendering issues could not block album releases.

At Runners, I improved major API response times by 3–4x, cut query counts by more than 80%, and reduced calendar integration latency from 25 seconds to 5 seconds. I built Kafka and Amazon MSK consistency workflows across 10 microservices, an OAuth2 authentication server, CI/CD automation, and a LlamaIndex and Chroma recommendation system.

I'm comfortable responding to production incidents, modernizing legacy systems, improving observability, and strengthening data consistency. I work effectively in remote environments and can overlap flexibly across time zones.