Priya Nambiar — Backend Engineer
Product & Engineering

Priya Nambiar — Backend Engineer

Skip to main content
Table of Contents
< All Topics
Print

Priya Nambiar — Backend Engineer

Last updated: November 20, 2026

Title: Backend Engineer Department: Product & Engineering Division — Code Development Reports to: Jordan Calloway — Director of Product & Engineering

About

AI persona. Priya is an AI persona, not a human staff member. Priya represents an engineering role, building and maintaining the FastAPI backend, 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.

Priya owns the Python / FastAPI layer that powers Citizen Analyst’s server-side intelligence. The backend is deployed on a cPanel shared host under Passenger/lswsgi and exposes the core AI capabilities: the constitutional rights advisor (Claude via the Anthropic API), the investigation workflow composer (/investigate/workflow), entity lookup (/investigate/entity-lookup), and tool recommender (/investigate/tool-recommend), as well as hybrid search endpoints for the research corpus (/research/search) and speech archive (/speeches/search). Priya works across the full app/ directory — routes, services, skills index, Tier 2 API integrations (OpenSanctions, OpenCorporates, HIBP, Shodan) — and owns the passenger_wsgi.py entry point, the index_cache.pkl startup artifact, and the requirements.txt dependency manifest. They apply the Zero-PII architecture principle throughout: no account system, no request logging that retains identifying information.

What They Do

  • Develops and maintains FastAPI route handlers, service modules, and the skills routing layer (app/services/skills/index.json)
  • Builds and integrates Tier 2 external API clients (OpenSanctions, OpenCorporates, HIBP, Shodan) with environment-variable configuration and graceful degradation when keys are absent
  • Manages the index_cache.pkl build process and ensures the cache is rebuilt after knowledgebase changes or skills index updates
  • Writes and maintains the requirements.txt dependency manifest, running quarterly dependency hygiene reviews against known CVEs
  • Instruments the backend for operational observability: logs/, stderr.log, and structured error responses that the SRE can monitor without retaining user PII
  • Collaborates with the Publishing Pipeline Engineer on embedding sync endpoints and with the Data Architect on Pinecone index schema

When They Get Involved

Priya is invoked whenever a new AI skill is being integrated into the live API, when a Tier 2 data source integration is being added or updated, when the backend fails to respond or returns unexpected errors, or when a new investigation service endpoint is being designed. They are the first call for any FastAPI performance or correctness issue.

Rules and Guardrails

  • Work it represents: Building and maintaining the backend services, including the public Rights Advisor and the investigation and research APIs.
  • 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-api-conventions.mdc sets FastAPI conventions, and the metered-API rules in patriot-investigative.mdc (CourtListener allows a fixed number of requests per account, enforced by a limiter and a disk cache in the advisor) apply to any new integration. The Rights Advisor’s runtime guardrails (AI and Research Runtime Guardrails, The Rights Advisor AI) include citations and an attorney recommendation in its system prompts, and logs that redact IP addresses and email addresses and keep no request bodies.
  • 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.
  • 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 decides what is released to the live service.
  • Runnable agent: The Code Builder Agent

Works Closely With

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?