PROJECT ASSIST
Interview companion
Understand · implement · explain · verify
Interview profile
Waiting for an interview profile…
Round focus sets the default assistance. Technical and leadership questions can alternate while your current problem and preparation remain available.
Capture health
Waiting for capture health
Network connectivity alone does not confirm that screen capture is working.
Interview background and reference documents
Waiting for session background…
Add your experience, role requirements, and prepared questions and answers to tailor guidance.
Round focus
Choose a round focus or let Auto follow the question.
Interview profile
Automatic guidance
Ready when you are
Guidance appears automatically as the agent hears the conversation and reads your screen.
Reasoning and follow-ups
Clarifying questions
Ask about missing requirements or context that could change your answer. Pinning holds this list while new questions arrive; guidance and implementation continue.
System design clarifying questions
Questions from the latest saved design. Design generation continues with stated assumptions while you clarify requirements. Pinning holds this list while new questions arrive.
Implementation plan
A brief first-person plan appears with the full implementation.
Full implementation
A complete solution appears automatically when the coding problem is clear.
Model suggestion. Automatic checks apply only to the exact implementation shown. Captured code and its test results remain separate below.
Held response
Stays here while guidance continues. Your microphone helps detect when you have covered the answer.
Waiting for a questionThe next detected question will appear here automatically.
Suggested STAR outline
Suggested wording
STAR feedback on your answer
Based on speech heard so far. Applies to behavioral and experience answers.
Tighter wording from your answer
Suggested wording and completion detection are model assessments; silence alone does not mean you have answered.
Observed whiteboard
Screen observations are separate from the proposed design. Unreadable or offscreen content is not assumed to exist.
Relevant source
Waiting for a relevant source…
Current problem
Waiting for a question…
Potential issue
No findings yet.
Model suggestions are unverified. Execution results appear below.
Round workspace
Every coding round
- Clarify requirements and label assumptions.
- Write correct, complete, readable code with useful comments.
- Choose suitable data structures and algorithms.
- Keep interfaces maintainable and responsibilities clear.
- Cover edge cases, error handling, and meaningful tests.
- Explain the approach, tradeoffs, time and space complexity.
Focus adds emphasis; every standard still applies. Implementation continues while clarifying questions are pending. Screen and audio observation require no coding-environment integration.
Proposed system design board
Project technical reference
Prepared source excerpts for ownership, decisions, rollout, failures and results.
Leadership Principles and your stories
Suggested associations need your review. Excerpts retain your original wording.
Choose a round focus or wait for a question to see its workspace.
GenAI engineering reference
Use technical reasoning for hypothetical questions. Personal claims require supporting experience.
- Define the customer problem, success criteria, and a non-AI baseline.
- Ground responses in authorized sources; preserve citations and handle missing or conflicting evidence.
- Evaluate factuality, retrieval quality, adversarial inputs, and regressions on representative cases.
- Enforce access controls before retrieval and tool actions; treat retrieved instructions as untrusted data.
- Discuss latency, token budgets, caching, fallbacks, monitoring, and human review.
- For audit evidence, preserve provenance and distinguish evidence retrieval from compliance conclusions.
Round practice
One prompt at a time. Record your own answer, then respond to follow-ups and changes.
Code and verification
No code selected. Configure a saved file on the Mac or paste browser code below.
Current captured code
Preflight and screen region
Preflight makes a small model request and sends a screen preview here. Verify that code text is readable and both audio meters respond.
Conversation
Help history
Activity and commands
Your experience · STAR
Your authored wording. Select one story; it stays expanded while AI guidance continues.
Waiting for session stories…
Stories across interviews
Suggested stories are not proof that you said them. Record or confirm usage to keep an accurate interview history.
Practice replay
Recurring practice topics
Delivery and performance
Capture computer delivery
Memory queue entries are not yet saved to disk. Pending-delivery counts and free-space readings may be cached while local storage is unavailable.
Performance measurements
Question to visible answer
Waiting for a live question.
Cross-device totals are estimates with clock uncertainty. Browser-only measurements begin when this page receives the question. History and replay do not add live samples.
Storage and backups
Export any sessions you want to keep before pruning. Retention removes whole inactive sessions, including their documents and commands. The selected session, agents active within five minutes, and sessions with pending commands are protected. Active sessions may temporarily exceed the event quota. Freed database pages are reused.
Database backups
Create a verified online SQLite snapshot on the dashboard server. Older backups are rotated within the configured count and byte budget. Backups contain private session information; session exports remain available above.