Mailflow Control Room
An operational dashboard for monitoring message pipelines, queue health, and delivery bottlenecks.
Built as a dense but readable operator surface with alerting, timeline views, and focused controls for intervention.
Personal RAG + Visual Builder
teemails
A public portfolio with a private engine behind it.
teemails.com pairs a sharply curated frontend with a fully local retrieval stack, letting visitors explore projects, experience, and contact paths without drifting into generic AI behavior.
Personal systems, local intelligence, and a portfolio that stays grounded in real work.
Built to stay specific
Answers stay anchored to the public record.
The chatbot only responds from curated resume, bio, project, and site materials. If a question falls outside that scope, it redirects people toward the right next step instead of improvising.
Selected work
Focused systems with clear operational edges.
Local RAG Workbench
A retrieval-first assistant stack for curated personal and project data running entirely on local infrastructure.
Combines document ingestion, embeddings, vector search, and local model inference with scope guardrails.
FAQ
The assistant is intentionally narrow.
What powers the chatbot?
A local FastAPI backend, PostgreSQL with pgvector, SentenceTransformers embeddings, and Ollama for final response generation.
What can visitors ask about?
Public information about background, project work, selected technical decisions, and contact or resume-related requests.
Does the assistant answer general knowledge questions?
No. It is designed to answer from curated teemails.com content only and otherwise redirect visitors to resume, projects, or contact actions.
Next step
Bring a concrete problem.
If the work sounds relevant, use the contact form and include enough context to make the next conversation useful.
Open contact form