Amna Shah

I’m a software engineering intern with hands-on experience improving semantic search, integrating LLM-powered query enrichment, and optimizing embedding pipelines. I enjoy turning messy user intent into reliable retrieval and measurable performance gains.

At Bazaar Technologies, I enhanced semantic search for a 10K+ SKU catalog by eliminating 90% of irrelevant results. I also integrated LLM calls to generate Roman Urdu variations, benchmarked search engines (OpenSearch, Algolia, Meilisearch), and monitored embedding pipelines in Apache Airflow with Kubernetes deployment verification via Argo CD.

My project work reflects the kind of engineering I like: end-to-end delivery and practical AI. For my final-year Opticode platform, I built a full-stack Dev–Learner matching experience using SonarQube static analysis and CodeBERT embeddings for clustering, and I’m comfortable working across Next.js, MongoDB, PostgreSQL, and containerized deployments.

I also develop secure, scalable backend services—building a NestJS + FastAPI food redistribution platform with Azure AI Vision, Azure Maps, and automated notifications through Serverless Azure Functions. Earlier, I designed a Node.js/Express.js mood tracker with Oracle SQL/PLSQL for user profiles and real-time community interactions.