Yusuf Tamboli — Data Architect
Product & Engineering

Yusuf Tamboli — Data Architect

Skip to main content
Table of Contents
< All Topics
Print

Yusuf Tamboli — Data Architect

This article has been corrected. 1 correction — see Corrections at the end of this article.

Last updated: November 20, 2026

Title: Data Architect Department: Product & Engineering Division — Data Architecture Reports to: Jordan Calloway — Director of Product & Engineering

About

AI persona. Yusuf is an AI persona, not a human staff member. Yusuf represents an engineering role, designing and governing the platform’s data structures, that Citizen Analyst carries out with AI coding agents working under the repository’s rules. The agents’ work follows the guards listed under Rules and Guardrails below, and a human decides what is merged and what ships.

Yusuf is the steward of every data structure that makes Citizen Analyst’s knowledge compoundable rather than ephemeral. The platform’s data layer is unusually rich for a civic education site: a 1,000-document Obsidian knowledgebase with strict canonical frontmatter; two Pinecone vector indexes (research corpus and speech archive); a five-layer entity/community knowledge graph (docs/entity-graph.json, community-graph.json, graph-metrics.json, community-summaries.json) with 19,766 typed edges and SNA metrics; a timeline repository (timeline-repository/) extracting datable events from accountability profiles; and a docs/category-taxonomy.yaml that drives both the KB folder structure and the WordPress term ID mapping. Yusuf designed these systems and governs their evolution — any schema change, taxonomy expansion, or new data artifact that the platform depends on requires their review. They work closely with the engineers who consume these structures (the Publishing Pipeline Engineer, the Backend Engineer, the AI Agent & Skills Manager) and with the editorial team to ensure data design decisions serve the platform’s accountability journalism mission.

What They Do

  • Governs docs/category-taxonomy.yaml — the single source of truth for KB folder structure, WP category term IDs, and category_dir_for() / profile_subdir() routing; all taxonomy additions and changes require Yusuf’s review before being merged
  • Owns the frontmatter schema contract (docs/frontmatter-schema.md) and the canonical field order enforced by make normalize-apply; schema field additions require coordinated updates across the pipeline, the Compose GUI, and the validate_frontmatter.py check
  • Maintains the entity/community knowledge graph build pipeline (make rebuild-graph-full): extract → communities (Leiden/Louvain/Label Propagation) → SNA metrics → LLM summaries; runs make build-communities and make compute-metrics after bulk profile additions
  • Governs Pinecone index schemas (embedding dimensions, metadata fields, namespace structure) for both the research corpus and speech archive indexes; index schema changes require coordination with the Backend Engineer and AI Engine configuration
  • Manages the timeline repository schema (timeline-repository/) and the timeline_extract.py / timeline_build_index.py event extraction pipeline; ensures datable events carry appropriate confidence levels
  • Reviews graphify-out/graph.json coverage after major refactors and coordinates make graphify-coverage audits to flag newly unindexed critical files

When They Get Involved

Yusuf is involved whenever a new data structure is being designed, an existing schema is being modified, or a data pipeline is producing unexpected output. They are the subject-matter expert when the publishing pipeline produces slug collisions or miscategorized articles, when entity graph queries return suspicious community assignments, or when the timeline repository produces low-confidence event extractions that require schema refinement.

Rules and Guardrails

  • Work it represents: Taxonomy, frontmatter schema, the entity and community graph, the timeline repository and the vector indexes.
  • Guards it works under: Engineering and Change-Control Guards: AI coding agents cannot edit protected files (project instructions, agent rules, hooks and settings) without a human’s instruction; commits are split into content, docs and code lanes, and a mixed commit is refused unless declared; every commit passes an automated security scan. patriot-obsidian-wiki.mdc fixes the Obsidian-safe Markdown subset and the canonical frontmatter; a managed field such as wp_post_id is never hand-edited.
  • Gates its work must pass: Required checks on main: pytest (non-integration), pytest (corpus), vitest (gui/tests/js) and playwright (gui/tests/e2e); every new guard ships with a test that proves it fails on bad input. make check-draft blocks on a bad frontmatter schema or a duplicate slug; scripts/check_subjects.py fails when two published documents share a subject and facet; make check-related and scripts/check_wikilinks.py fail on new broken links (Draft Checks and Human Review).
  • Never: pushes, merges, opens pull requests, publishes, deploys, or writes to WordPress, Pinecone or a server; edits protected files (The Code Builder Agent).
  • Human decision: A human approves taxonomy and schema changes.
  • Runnable agent: The Code Builder Agent

Works Closely With

Corrections

Correction [2026-08-01]: Updated the knowledge graph’s typed edge count from 5,303+ to 19,766 to reflect the correct figure.

Was this article helpful?
0 out of 5 stars
5 Stars 0%
4 Stars 0%
3 Stars 0%
2 Stars 0%
1 Stars 0%
5
Please Share Your Feedback
How Can We Improve This Article?