> ## Documentation Index > Fetch the complete documentation index at: https://docs.witting.ai/llms.txt > Use this file to discover all available pages before exploring further. # Subject Matter Interviews > Set up Vocalis AI Actors for technical and domain-specific interviews. Includes a sample Senior Backend Engineer AI Actor with Purpose, Insights, and evaluation criteria. Subject matter interviews check whether a candidate has the specific technical skills, domain knowledge, or professional expertise the role requires. ## When to use a subject matter interview AI Actor * **Technical roles** — Software engineers, data scientists, DevOps engineers * **Domain-specific roles** — Finance analysts, healthcare professionals, legal specialists * **Any role requiring verifiable expertise** that can be assessed through conversation ## Tailoring questions to the role Subject matter interviews need to be customized for each role. The Purpose should include: 1. **Role context** — What the position involves day-to-day 2. **Technical questions** — Specific to the skills required 3. **Scenario-based questions** — "How would you handle..." situations 4. **Depth probes** — Follow-up patterns to tell the difference between surface knowledge and real expertise ## Sample AI Actor: Senior Backend Engineer ### Step 1: Actor basics | Setting | Value | | ------- | -------------------------------------- | | Name | Technical Interview - Backend Engineer | | Voice | Professional, confident tone | ### Step 2: Purpose ``` You are conducting a 30-minute technical interview for a Senior Backend Engineer position at {{company_name}}. ## Your role You are Jordan Lee, a Senior Engineering Manager. You're evaluating whether this candidate has the depth of backend experience needed for a senior role on your team. ## Context This interview focuses on technical depth. The role involves building and maintaining distributed systems processing 10,000+ requests/second. ## Interview structure ### Part 1: Architecture (10 min) Ask the candidate to describe the architecture of a system they built or significantly contributed to. Probe on: - Scale and traffic patterns - Database choices and why - How they handled failure modes - What they'd change if they rebuilt it ### Part 2: Technical deep-dive (10 min) Choose ONE based on their background: If backend/systems focused: - "Walk me through how you'd design a rate limiter for an API serving 10,000 req/s. What data structures would you use?" - Follow up on edge cases, distributed scenarios, monitoring If data focused: - "How would you design a pipeline that processes 1M events/day with exactly-once delivery guarantees?" - Follow up on failure handling, replay, monitoring ### Part 3: Problem-solving (10 min) - "Tell me about the hardest bug you've debugged in production. Walk me through your process." - Follow up: What tools did you use? How did you prevent it from recurring? ## Evaluation criteria Rate each area (1-5): - System design thinking - Technical depth in their primary area - Problem-solving approach - Communication of technical concepts Provide overall recommendation: Strong Yes / Yes / Maybe / No Include specific technical strengths and gaps observed. ``` ### Step 3: Insights | Insight | Type | Purpose | | ------------------------------- | ----- | ---------------------------------------- | | Architecture discussion summary | Text | Key points from their system description | | Scale experience | Text | Largest systems they've worked on | | Primary technical strengths | Text | Areas of deepest expertise | | Technical gaps identified | Text | Areas where knowledge was shallow | | System design rating | Score | 1-5 | | Technical depth rating | Score | 1-5 | | Problem-solving rating | Score | 1-5 | | Communication rating | Score | 1-5 | | Overall recommendation | Text | Strong Yes / Yes / Maybe / No with notes | ## Writing good evaluation criteria For subject matter interviews, clear evaluation criteria are essential: * **Be specific about what "good" looks like** — Instead of "strong technical skills", try "can explain trade-offs between SQL and NoSQL for their use case." * **Include red flags** — "If the candidate cannot explain why they chose a particular database, rate Technical Depth as 2 or below." * **Check with the hiring manager** — Have them review the Purpose and confirm the questions match what they would ask in person. ## Adapting for other roles The same pattern works for any domain. Adjust the questions and evaluation criteria: | Role | Question focus | Example scenario | | ----------------- | ----------------------------------------------- | ------------------------------------------------------------------------------------------------------- | | Data Scientist | Statistical methods, ML pipeline design | "How would you detect and handle data drift in a production model?" | | DevOps Engineer | Infrastructure, CI/CD, incident response | "Walk me through your approach to a cascading failure in a microservices architecture." | | Product Manager | Prioritization, metrics, stakeholder management | "How would you decide between two features with equal customer demand but different engineering costs?" | | Financial Analyst | Modeling, valuation, risk assessment | "Walk me through how you'd build a DCF model for a SaaS company." | ## Next steps Add behavioral interviews to your pipeline. Improve your technical interview instructions.