Skip to content
SphinqueSphinque

The product

Relevant work. Inspectable engineering judgment.

A role-aligned work sample, a technical follow-up, and an evidence brief. Built to help teams replace or compress the first technical screen.

Explore the evidence

01 / Role alignment

Start with the job. Not a blank prompt.

A job description starts the conversation. Confirm the work, stack, seniority, and critical competencies before choosing the assessment.

  • Agree on the engineering decisions that matter.
  • Choose a supported, relevant task pattern.
  • Make the evaluation boundaries explicit.
  • Apply a consistent standard across candidates.

Extend a multi-tenant orders API

Investigate an authorization gap, implement a scoped change, and test cross-tenant behavior.

Stack
Python · FastAPI · SQLAlchemy
Evaluate
Access control · Verification
Not assessed
People leadership · Front-end design
  • Debugging
  • Data boundaries
  • Failure modes

Example configurations. Supported today: backend roles on Java and Python, and mid-level frontend roles on React. Confirm role and stack fit in a walkthrough.

02 / One connected evaluation

The work is the starting point. The evidence is what you keep.

Preserve the context between implementation, verification, and technical follow-up.

Engineering evidence brief

Sample candidate · Senior Backend Engineer

Human review
Verification judgment

Recognized a missing isolation test. Added coverage.

02 evidence items
Technical understanding

Connected access control to the query boundary.

03 evidence items
Open question: operational resilience

Retry and failure-recovery reasoning needs follow-up.

Review required
Illustrative report · Not a hiring decisionExplore the brief
03 /

A realistic codebase

Inspect how a candidate navigates existing code, scopes a change, and reasons about constraints. The goal is a bounded engineering work sample, not unpaid production work.
04 /

AI-enabled work

Make permitted AI use explicit. Examine the candidate’s decisions about generated code without equating prompt volume, typing speed, or tool preference with competence.
05 /

Technical follow-up

Ground the discussion in consequential choices from the task: an omitted test, an access boundary, a failure mode, or a changed requirement.
06 /

A useful next conversation

Present demonstrated strengths, supporting evidence, limits, and follow-up questions. Give human reviewers room to disagree with the interpretation.

A better first screen. A better next conversation.

See what engineering evidence could look like for your next hire.

Book a walkthrough