Lynn.

Extended an open-source AI job-search framework into a local-first application workspace with multi-source ingestion, de-duplication, freshness checks, fit ranking, and lifecycle tracking.

Built agent-assisted evaluation and drafter-reviewer workflows for tailored CVs, cover letters, and interview preparation, with a private dashboard, local document compilation, and privacy-preserving document handling.

Implemented a modular ingestion and ranking pipeline with caching and incremental updates to keep job data fresh and relevant.

Designed and built a responsive English/Chinese website with Contentful-backed content management, role-aware administration, and a controlled publishing workflow.

Implemented real-time flavour, ordering, visibility, and menu management, plus internationalised routing, SEO metadata, sitemap generation, and production deployment.

Set up CI/CD with GitHub Actions and edge deployment on Vercel and Cloudflare for fast, reliable releases.

Built an end-to-end RAG pipeline for policies, announcements, and service procedures, covering parsing, chunking, embeddings, semantic retrieval, and answer generation.

Designed grounded prompts and retrieval controls to reduce hallucinations and improve answer accuracy and explainability.

Built document ingestion and vector indexing with Redis Stack and object storage, with evaluation to measure retrieval quality and answer accuracy.