At Zeus Learning, I designed and implemented a Go-based AWS migration pipeline that processed third-party API data under a shared rate limit. Its initial run migrated 130,000+ records, and I improved PostgreSQL JSON transformation and Elasticsearch publishing times through indexing, paginated reads, and bulk APIs.
I also designed an AI-powered question and metadata generation pipeline using Go, Python, and Amazon Bedrock. With retrieval-augmented generation, automated evaluation, and human approval before publishing, the workflow generated 100,000+ questions in 15 days.
I led a three-person team developing a Python-based bidirectional CMS migration tool, owning its core architecture and implementation. Earlier at Zeus Learning, I built a job-portal capstone during my internship, and I also built a Java TCP server implementing the RESP protocol.

