The Translation
Federal titles on the left, plain language titles on the right. Hover to see connections. Click any title to see individual postings.
How this works
The Problem
Federal job announcements use titles like "Management and Program Analyst (GS-0343)" or "IT Specialist (INFOSEC)." A private-sector jobseeker has no idea what these roles involve. Someone who'd be a perfect fit for a "Cybersecurity Compliance Manager" role would never think to search for "IT Specialist (PLCYPLN)."
In September 2025, OPM issued job titling guidance under Executive Order 14170, directing agencies to use titles that "resonate with jobseekers" and "align with private-sector terminology."
The Pipeline
Read the duties
We pull job postings from USAJobs for a target occupational series, extracting the duties, qualifications, and summary text that describe what the person does day to day.
Propose real titles
An LLM reads each posting's duties—without seeing the current title—and proposes titles you'd see on LinkedIn. The prompt explicitly avoids vague catch-alls like "Program Manager" or "Management Analyst."
Consolidate & deduplicate
Raw proposals get programmatically deduplicated—normalizing prefixes like "Senior" and "Lead," merging near-duplicates, and filtering out one-off suggestions. Each surviving title then gets individually validated by the LLM: "Would a recruiter post this on Indeed? Would a jobseeker know what this person does?"
Validate against OPM guidance
Titles are checked against OPM's September 2025 plain-language titling guidance and EO 14170. Titles that are still too generic get replaced with more specific alternatives suggested by the LLM, then enriched with plain-English descriptions.
Classify every posting
Each posting gets matched to the best-fitting title with a 1–5 fit score. The LLM classifies based on duties alone—it doesn't see the current title. If a posting's original title passes a plain-language check, it's included as one of the options. If nothing fits well, the LLM proposes a new title on the spot.
Measure coverage
We compute coverage rates, per-title breakdowns, and flag potential issues—titles that are too broad, too narrow, or poorly aligned—so the title set can be iteratively refined.
What you're seeing
The graph above shows how federal titles (left) map to plain language titles (right). One federal title often covers several distinct roles. Hover over any title to see its connections, or click to drill into individual postings with the classification reasoning.
This is a starting point, not an exhaustive menu. Some postings already have clear, descriptive titles that don't need changing. The goal isn't to force every job into a new label—it's to show where the current titles don't capture what people do, and suggest alternatives for those cases. Hiring managers know their roles best; these suggestions are meant to give them a head start.