Projects
This page separates the flagship build from reference patterns. The goal is to be honest about what is being built, what each project proves, and which artifacts should be added next.
Flagship build
A flagship build track for turning an AI agent demo into a controlled, observable, testable product system.
Full-stack scope: Frontend workflow UI, backend orchestration, retrieval layer, tool execution gateway, evaluation runner, and trace dashboard.
What it proves: How I think across the full stack: UX, APIs, state, data, model behavior, evals, and deployment readiness.
Reference patterns
A practical RAG product pattern for internal knowledge workflows where users need sourced, reviewable answers.
Full-stack scope: Document ingestion, metadata, vector retrieval, answer generation, citations, feedback capture, and admin review.
What it proves: Ability to design AI features as product workflows, not just API calls to a model.
A dashboard pattern for comparing prompts, tracking model outputs, reviewing traces, and catching regressions before release.
Full-stack scope: Evaluation datasets, rubric scoring, run history, model comparison, trace review, and release readiness signals.
What it proves: Understanding that production AI quality depends on measurement, not vibes.
A teaching and mentoring product direction for turning AI engineering concepts into hands-on implementation exercises.
Full-stack scope: Learning paths, project templates, student notebooks, code walkthroughs, workshop pages, and feedback loops.
What it proves: Ability to explain complex AI engineering clearly and convert knowledge into structured learning experiences.
The next quality jump will come from artifacts, not more pages. These are the proof objects that should be added as the flagship matures.
