New opportunity

AI Strategist

Centh AI · New York, United States

About this role

Centh AI is hiring an AI Strategist to turn recruiting and operations processes into agent-powered workflows built on customers’ applicant tracking systems. The role sits between customers, product and engineering: the strategist works with users to understand existing processes, defines where agents act and humans make decisions, and partners with engineers to put workflows into production. The position is listed as full-time in New York, NY.

Responsibilities

Interview recruiters, hiring managers and operations leaders about sourcing, screening, compliance and follow-up. Convert observations into structured workflows, diagrams and requirements engineers can use.
Design multi-step agent flows, including agent responsibilities, human approval or intervention points, and measures of success. Write prompts, tool-calling patterns and guardrails for repeatable behavior.
Define quality standards for outputs such as shortlists, screening summaries and submissions. Establish qualitative and quantitative evaluation loops and use the results to improve workflows.
Support early customer pilots through onboarding, workflow setup, feedback collection and iteration. Contribute to demos, playbooks and explanations of the workflows.

Qualifications

Experience translating complex real-world processes into structured flows, documents, playbooks or product specifications, and working directly with operators to turn their input into clear requirements.
Ability to break vague operational problems into concrete workflows and experiments. Hands-on work with LLMs or agents, or self-directed projects involving the building and testing of AI-driven flows.
Comfort using both narratives and metrics, including time-to-fill, conversion, redeployment and revenue per recruiter; strong written communication; and the ability to collaborate with engineers on implementation details. Interest in an early-stage team with direct feedback and substantial ownership.
Experience in staffing, recruiting, HR technology or similar operations is a plus, not a requirement.

To apply, email a resume or portfolio and brief introduction to jobs@centh.ai, including the role title in the subject.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

AI agentsLLMsWorkflow designProcess mappingPrompt writingTool callingGuardrail designQualitative evaluationQuantitative evaluationRequirements gatheringStakeholder interviewsRecruiting operationsApplicant tracking systemsWritten communication

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Map recruiting operations into buildable requirements
The role interviews recruiters, hiring managers and operations leaders about sourcing, screening, compliance and follow-up, then turns observations into workflows and diagrams engineers can use.
Design agent and human handoffs
The strategist specifies multi-step flows on top of customers’ applicant tracking systems, including what agents do, where people approve or intervene, and how success is measured.
Make agent behavior repeatable
The role calls for prompts, tool-calling patterns and guardrails, developed with engineers to move workflows into production rather than leave them as one-off prompts.
Define quality and improve it through evaluation
The strategist sets standards for outputs such as shortlists and screening summaries, then uses lightweight qualitative and quantitative evaluation loops to refine workflows.
Turn early pilots into usable workflows
Customer onboarding, workflow setup, feedback collection and iteration are part of the job, alongside demos and playbooks that explain what the workflows do and why they matter.

Likely challenges to think through

Resolving inconsistent accounts of the same process
Likely preparation challenge, not a confirmed problem at Centh: recruiters, hiring managers and operations leaders may describe different versions of a workflow. Prepare an example of how to reconcile interviews with observations, document exceptions and turn an ambiguous process into testable requirements.
Balancing automation with human judgment
Likely preparation challenge: more agent autonomy may speed a flow but make sensitive or uncertain decisions harder to review. Prepare to explain how you would choose approval points, intervention paths and guardrails for a screening or compliance-related flow without assuming every step should be automated.
Proving value during an evolving pilot
Likely preparation challenge: early feedback may be qualitative or limited, while the role also calls for measurable success. Prepare an evaluation approach that pairs output review with a relevant operational metric, states what would trigger a workflow change and avoids claiming improvement before it is measured.

Practice your answers

Technical practice

Suggested practice: How would you map a recruiter’s current sourcing-to-follow-up process?

What to coverThe role calls for interviews with recruiters, hiring managers and operations leaders, followed by workflows engineers can build from. Suggested approach: show how you would distinguish observed steps from assumptions and document handoffs, decisions and exceptions.

Suggested practice: How would you turn an inconsistent ATS-based process into implementation requirements?

What to coverThe posting places agent-powered workflows on top of customers’ applicant tracking systems. Suggested approach: define the inputs, workflow steps, expected outputs and unresolved questions you would bring to engineering, without assuming a particular ATS.

Suggested practice: Where should agents act, and where should humans decide?

What to coverThe role explicitly involves assigning agent responsibilities and human approval or intervention points. Suggested approach: walk through a screening or submission flow and explain how you would choose those boundaries and measure whether they work.

Suggested practice: How would you design prompts and tool calls for a repeatable multi-step flow?

What to coverWriting prompts and tool-calling patterns is a stated responsibility. Suggested approach: explain what each step needs to receive and produce, and how you would test whether the flow behaves consistently across cases.

Suggested practice: How would you troubleshoot an agent flow that produces an unreliable screening summary?

What to coverThe posting asks for guardrails and robust, repeatable behavior rather than a one-off prompt. Suggested approach: isolate whether the issue comes from inputs, tool results, instructions or handoffs, then describe a targeted change and retest.

Suggested practice: What makes a shortlist, screening summary or submission good?

What to coverDefining quality standards for these outputs is part of the role. Suggested approach: propose criteria that users can judge consistently, including usefulness and clarity, while noting where human judgment remains necessary.

Suggested practice: How would you set up a lightweight evaluation loop?

What to coverThe role calls for both qualitative and quantitative evaluation and iteration. Suggested approach: describe how you would collect user feedback, track a small set of output-quality measures and decide what to change next.

Suggested practice: How would you test whether a workflow improves a vague operational problem?

What to coverThe posting names problems such as slow or inconsistent processes and metrics including time-to-fill, conversion, redeployment and revenue per recruiter. Suggested approach: choose a relevant baseline and experiment, then explain the trade-off between a faster workflow and output quality.

Suggested practice: What would you give engineers to move a workflow toward production?

What to coverThe strategist is expected to turn observations into diagrams and requirements and work alongside engineers on implementation details. Suggested approach: describe a handoff that makes the flow, agent and human responsibilities, success measures and open decisions clear.

Suggested practice: How would you use feedback from an early customer pilot?

What to coverThe role includes onboarding, workflow setup, feedback collection and iteration. Suggested approach: explain how you would separate setup problems from workflow-design problems, prioritize changes and communicate what the pilot has and has not shown.

Behavioural practice

Suggested practice: Tell me about turning operators’ messy process into a clear plan.

What to coverUse a real situation to explain whom you spoke with, what you observed and how you turned conflicting or incomplete input into a usable artifact. Close with the result and what remained uncertain.

Suggested practice: Describe a time you worked with engineers to resolve an implementation trade-off.

What to coverGive the situation, your role and the actions you took to make requirements or constraints explicit. Explain the decision reached and its observable result rather than claiming a perfect outcome.

Suggested practice: Tell me about a pilot or early rollout you changed after user feedback.

What to coverSet out the initial goal, how you collected and interpreted feedback, and what you changed. Describe the result and how you knew whether the change helped.

Suggested practice: Describe a time you took ownership of an ambiguous problem.

What to coverThe posting emphasizes breaking vague problems into experiments and working with substantial ownership. Use a real situation/action/result example that shows how you set a scope, chose a first test and adjusted course.

Suggested practice: Tell me about receiving direct feedback on work you had already shared.

What to coverThe posting describes an early-stage team with direct feedback and calls for strong written communication. Explain the situation, how you responded and what changed in the work or its outcome.

Sharpen your knowledge

Topics to prepare

Recruiting workflows and ATS context
The role maps sourcing, screening, compliance and follow-up into workflows on top of customers’ applicant tracking systems (ATS). Revision exercise: sketch a hypothetical candidate journey through those stages, marking the information and decisions needed at each step. The scenario is practice, not a stated Centh workflow.
Agent and human decision boundaries
The posting calls for multi-step flows that specify agent responsibilities and human approval or intervention points. Revision exercise: take one step in your hypothetical workflow and explain what an agent could do, what a person must decide, and when the agent should stop or escalate.
Output quality and evaluation
The role defines good shortlists, screening summaries and submissions, then uses qualitative and quantitative evaluation to improve them. Revision exercise: draft criteria for a screening summary, including one review question and one measurable check.
Operational metrics and experiments
The posting names time-to-fill, conversion, redeployment and revenue per recruiter, and asks for concrete experiments from vague problems. Revision exercise: choose a hypothetical slow screening process, propose one workflow change, and state which metric and operator feedback would help assess it.
Pilot iteration
Early customer pilots involve onboarding, workflow setup, feedback collection and iteration. Revision exercise: outline what you would learn before setup, what feedback you would request after use, and how you would decide on one revision. This outline is preparation, not a claimed Centh pilot process.

Skills to revise

Interview operators and map a process
The role requires talking with recruiters, hiring managers and operations leaders and converting observations into diagrams. Practice a short interview about a hypothetical screening process, then draw its steps, exceptions and decision points. Compare the diagram with your notes to spot assumptions.
Write engineer-ready requirements
The posting asks for clear requirements and close collaboration with engineers. Turn one mapped workflow step into a brief specification covering its input, expected output, decision owner, exception path and success measure; mark any unanswered questions for the operator.
Design and test an agent flow
Hands-on LLM or agent work is requested, including prompts, tool-calling patterns and guardrails. As a practice suggestion, draft a prompt and a tool-call sequence for one ATS-backed step, specify an approval boundary, and test how the flow responds to incomplete information. No particular implementation tool is specified.
Evaluate sample outputs
The role calls for lightweight qualitative and quantitative evaluation loops. Create two hypothetical screening summaries, apply the same quality criteria to both, record the results, and revise the prompt or workflow to address a specific failure.
Communicate a pilot workflow
The posting includes customer feedback, demos and playbooks. Practice writing a one-page explanation of a proposed flow: what operators do, what agents do, where approvals occur, how success is checked, and what feedback would prompt a change.

Your preparation checklist

Choose a workflow walkthrough
Prepare a concise example of turning a messy operational process into a flow, specification or playbook. Show the operator input, decisions made, engineer-ready output and result; label any hypothetical example clearly.
Practice an operator discovery interview
Draft questions for recruiters, hiring managers and operations leaders about sourcing, screening, compliance and follow-up. Prepare a sample notes-to-requirements summary showing how you would resolve an ambiguous request.
Sketch an ATS-based agent flow
As a practice exercise, diagram a multi-step recruiting workflow on top of an ATS. Mark each agent’s task, human approval or intervention points, and what engineers would need to implement it.
Prepare a prompt and testing example
Bring a prompt, proposed tool-calling pattern and guardrails for one step in your example flow. Explain how you tested or would test repeatability and handle a failed output.
Define output quality and evaluation
Create a sample rubric for a shortlist, screening summary or submission. Pair qualitative review with a quantitative measure, explaining how findings would drive an iteration; the posting cites time-to-fill, conversion, redeployment and revenue per recruiter as relevant metrics.
Outline an early-pilot plan
Draft a one-page plan covering customer onboarding, workflow setup, feedback collection and iteration. Include a brief demo narrative or playbook excerpt that explains the flow to users.
Prepare questions for the team
Ask how customer workflows are selected, how strategists and engineers divide implementation work, where customers retain decisions, and how early pilots define useful outcomes.
Assemble the application materials
Prepare a resume or portfolio and brief introduction tailored to the Founding AI Strategist role. The posting asks applicants to email them to jobs@centh.ai with the role title in the subject.

Availability last checked: Nov 3, 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in United States. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…