The Mechanism

From your engineering workflow to a personalized upskilling path in days.

Tunlai reads your engineering workflow signals — not your engineers. Three steps, each building on the last. The competency graph emerges from the data your team already produces as a byproduct of normal work. No surveys. No self-assessments. No behavior change required from engineers.

Request Access
Step 01 — Connect

Connect your engineering data sources.

Authorize read-only OAuth connections to GitHub, GitLab, Jira, Confluence, Slack, and PagerDuty. Tunlai never stores your source code — it reads metadata signals: PR reviews, ticket assignment patterns, incident involvement, and documentation edits.

  • Setup takes under 30 minutes
  • Read-only OAuth scopes — no write access ever
  • Source code is never transmitted or stored
  • Revoke access at any time from your integration settings
Abstract visual representing data connection and integration between engineering tools
Step 02 — Analyze

The competency graph builds from your workflow signals.

Tunlai processes your workflow signals to construct a live competency graph: what your codebase demands, which engineers cover each domain, where gaps exist, and which gaps correlate with incidents or escalation patterns. Skill nodes map to the domains your infrastructure actually operates in — not a generic taxonomy. The graph isn't built once and frozen; it updates as new PRs are reviewed, incidents are resolved, and tickets reassigned.

  • PR review comment patterns map reviewer vs. contributor skill asymmetry
  • Incident involvement records link to specific skill domain exposure
  • Ticket reassignment chains signal undocumented knowledge gaps
  • Documentation edit history maps knowledge ownership to individuals
  • Graph recomputes continuously as new workflow signals arrive
Abstract visualization of data analysis and pattern recognition across a competency graph
Step 03 — Path

Personalized learning paths, ranked by urgency.

Each engineer receives a ranked list of learning priorities scoped to the skills their specific codebase context requires. Urgency ordering puts incident-correlated gaps first — the gaps currently creating production risk — before moving to project-relevant gaps and general coverage improvements. Paths connect to the learning content your team already uses: Tunlai is a path engine, not a content library.

  • One path per engineer, updated as the competency graph evolves
  • Urgency ranking: incident-correlated gaps surface before general coverage gaps
  • Works with your existing content — Pluralsight, O'Reilly, internal wikis
  • L&D manager and Engineering VP share the same progress view in real time
  • Path reranks automatically when new incidents or project context arrives
Abstract visualization of a personalized learning path being generated

Your code never leaves your infrastructure.

Three things engineers ask us every time. Here are straight answers.

Read-only access

Tunlai requests the minimum OAuth scopes required to read metadata signals. No push, no write, no admin. You can see exactly what we access before authorizing.

No code storage

Source code is never transmitted to Tunlai servers. We process metadata: file paths, review comment patterns, ticket IDs. Not your actual code content.

Full audit log

Every data read is logged. Your security team can review what was accessed, when, and for what purpose. Revoke access at any time from your settings.

Full security details

Curious how it maps your team?

Connect your first integration and see your initial competency graph within days.