Know how an engineer builds a system.

Truen turns selected engineering work into a readable hiring profile: what they built, how they reasoned, where the evidence is thin, and what to probe next.

Create account Inspect an illustrative profile

The resume gets the interview. It can't explain the work.

Resume

Polished history, but no trail of how a decision was made.

Take-home

Finished output, but little proof of iteration or ownership.

Interview

A conversation, but not the work that preceded it.

The question is not whether someone used AI. It is whether they still owned the thinking.

One evaluation layer. Three ways in.

Start with an open role, a person you already found, or an opt-in directory. The destination is the same: specific proof for a better hiring conversation.

Post a job.

Receive candidate-consented profiles with the context needed to make a focused next decision.

Create an account

Bring someone you found.

Submit an attested public GitHub profile and build a report around their owned public work.

Open your directory.

Receive opt-in candidate submissions with a frozen profile projection, ready for review.

An illustrative profile, end to end.

This is a hand-authored example of the report structure. It is not a customer profile or a generated production result.

Read-only Truen Profile

Maya Chen

Founding full-stack engineer · Toronto, CA · Generated Mar 14, 2026

Overall signal

Real systems depth; ownership under pressure still unverified.

Maya shows stronger-than-typical evidence for early-career systems thinking: she decomposes failure modes, writes validation around edge cases, and revises AI-assisted code before it lands. The profile does not claim production impact beyond observable repository signals.

Evidence inspected

maya/event-ledger

318 authored commits, queue replay service, CI and integration tests detected.

maya/clinic-scheduler

Multi-role app with auth flows, migrations, and documented scheduling tradeoffs.

Trajectory

Accelerating

Later commits show more pre-code planning, smaller changes, and better tests than the first project slice.

Startup Fit

Builder-operator

Confidence: High

Maya shows credible end-to-end shipping behavior and returns to live projects after the first release. The strongest evidence is concentrated in two repositories, so this remains descriptive rather than predictive.

Shipping & finishing

Strong

Features reach deployed, testable states.

End-to-end range

Strong

Work spans data model, backend, and interface.

Iteration after launch

Strong

Returns to address externally visible defects.

Pragmatism

Moderate

Makes clear reliability tradeoffs on core paths.

Corroborated experience

Insufficient evidence

Solo work gives little team-context evidence.

AI collaboration

Accelerator

High confidence

Claude Code sessions on her connected repos show verified, well-directed work. She runs tests after AI edits, states decisions before delegating, and the same files show up in commits. Not inferred from GitHub alone.

Measured signals

Sessions

2 repos · 34-day span

72%

Verified after edits

tests or builds run on AI changes

61%

Planned before editing

plan mode before execute

Prompts per chat

1.4 chats per active day

Corroborated by commit history: session activity overlaps authored commit days on event-ledger.

Verification & error recovery
Runs the test suite after AI edits and pastes real failures back with framing. AI output does not ship unchecked.
High confidence
Ownership & direction
States decisions before delegating ("keep the ledger append-only, so…") and interrupts to redirect when the agent drifts.
High confidence
Planning & scoping
Non-trivial work starts in plan mode with constraints and file paths up front, not a one-line generate request.
High confidence
Evidence-driven iteration
Debugging loops supply new evidence each round — reproduction steps, narrowed stack traces — instead of "still broken".
Moderate confidence
Tooling investment
Invokes Supabase MCP for schema checks and Playwright for test runs. Corroborating signal, not the verdict.
Moderate confidence

Tooling observed: MCP: supabase · playwright · 3 custom skills · project CLAUDE.md

Qualification Fit · Founding engineer

Strong fit for 2 of 3 role priorities

The evidence supports full-stack ownership and reliability-minded systems work. Probe product judgment in collaborative settings before treating the fit as complete.

Own end-to-end features

Strong fit

Schema, APIs, UI, and follow-up fixes appear in the same authored windows.

Build reliable systems

Strong fit

Replay handling, regression tests, and CI are visible in the event-ledger work.

Operate with a small team

Partial fit

The work is largely solo, leaving collaboration behavior under-observed.

What they built

One card per analyzed repository. Shipped facts are measured from the repository and its deployment, not self-reported.

maya/event-ledger

Show details

Visit deployment

Deployed · LiveCI passingMaintained 118dTypeScript · Postgres · Queues

An event replay service that protects downstream work from duplicate delivery. Maya built the recovery path and validation around the replay boundary rather than treating the queue as a black box.

Models idempotency before side effects.
src/replay/consumer.ts and duplicate-delivery tests
Changes the design after failure-mode testing.
Replay cursor fix sequence and regression test

The reliability work is convincing, but external production usage is not verified.

maya/clinic-scheduler

Show details

Visit deployment

Deployed · LiveCI passingMaintained 84dTypeScript · Next.js · Postgres

A multi-role scheduling application with authenticated booking workflows. The repo shows Maya carrying a product slice through migrations, server actions, and user-facing scheduling constraints.

Owns one feature across the full stack.
Migrations, actions, and scheduling UI in the same commit window
Closes a booking edge case with a regression test.
fix double-booked slot commit sequence

Administrative surfaces are less resilient than the scheduling core.

Work experience

Self-reported · Unverified

Written by the candidate and frozen when this profile was shared. Not evidence-backed — where the analyzed repositories corroborate a claim, the evidence-backed sections say so.

Software engineering intern · Northline HealthMay 2025 — Aug 2025
Candidate-provided experience. Not used as evidence for this profile.

Flags

Limited team-review evidence

medium

Most work is solo; little evidence of review cycles or collaboration responses.

Concentrated in two repositories

low

Signal comes from two repos, so breadth is not yet established.

Impact proxies

Deployment URL detected for clinic scheduler, but no production usage metrics are claimed.
Issue history shows two externally reported bugs closed with follow-up tests.
Instrumentation config is absent, lowering confidence in runtime maturity.

Risks to probe

Needs live probing on ownership of AI-assisted refactors.
Admin and reporting surfaces show less resilience than core pipelines.
Most collaboration evidence comes from solo work, not team review.

Limitations

This profile is based on selected repositories only.
No private production logs, employer references, or live pair-programming evidence were used.
AI collaboration is read from connected tool sessions; repos without linked sessions are not assessed for AI style.

Interview follow-up

Walk through the replay cursor bug and explain what invariant the final design protects.

Show one AI-assisted change you rejected or rewrote, and explain the difference.

Your work. Your consent. A better application.

Candidates choose the repositories that represent their work, explicitly generate a private profile, and consent when using it for an application or directory submission. Claude Code pairing is optional.

Build your profile

Trust needs limits in writing.

No black-box rank

Truen supports a hiring decision. It does not make one.

Visible uncertainty

Thin or missing evidence stays insufficient instead of becoming a flattering conclusion.

Private by default

Applications and directory submissions share frozen profile projections, not a live candidate profile.

Bounded inputs

Candidates select repositories. Claude Code signals are optional and collected through a local consented flow.

Make the next engineering interview count.

Create an employer workspace, publish a role, and start reviewing work with its context intact.

Create account