Customer Support Task Author (AI Training)

Alignerr; Greater London; Remote; GBP 34,000 - 83,000; Full time

Job summary

Alignerr is building a dataset of expert tasks to train and evaluate AI agents in real enterprise support scenarios. As a Task Author for Customer Support / Success, you will design end-to-end triage tasks—reading tickets, checking configurations, consulting runbooks, and choosing between fix and escalation—reflecting real judgments under ambiguity.

This fully remote, flexible hourly contract is open to seasoned support engineers and customer success professionals based in London and across the

Qualifications

• Experienced support engineer, technical support specialist, or customer success engineer with hands-on product support experience.
• Deep familiarity with real product failure modes, escalation paths, and support tooling (ticketing systems, internal knowledge bases, runbooks).
• Ability to reason through ambiguous support scenarios and articulate why a given routing decision is correct.
• Comfortable defining verifiable correctness criteria — not just "good customer service" but an objectively defensible resolution or escalation decision.
• Strong written communication; tasks and rubrics must be precise enough to be evaluated without interpretation.

Responsibilities

• Author realistic support task prompts: customer issue triage, account configuration checks, runbook lookup, and fix-versus-escalation decisions.
• Write scoring rubrics that define the correct routing decision, resolution path, or escalation criteria in terms that can be objectively checked.
• Set up task environments with ticket queues, account records, runbooks, and system states that replicate a genuine support scenario.
• Solve each task yourself to confirm the scenario is realistic, the correct answer is clear, and the rubric captures the right reasoning.
• Calibrate task difficulty — adjust ticket ambiguity, account configuration complexity, or runbook depth until the task challenges the model appropriately.
• Review and correct AI-drafted task prompts or rubrics when provided.

Job description

About The Role

Alignerr is building a dataset of expert tasks that train and evaluate advanced AI agents on real enterprise work. As a Task Author for the Customer Support / Success role, you will design and calibrate end‑to‑end support triage tasks — reading a ticket, checking account configuration, searching runbooks, and deciding between fix and escalation — that reflect the real judgment calls experienced support professionals make under ambiguity.

This is a fully remote, flexible contract role open to seasoned support engineers and customer success professionals based in London and across the UK who want to shape the future of AI on their own schedule.

• Organization: Alignerr
• Type: Hourly Contract
• Location: Remote
• Commitment: 10–40 hours/week

Key Responsibilities

• Author realistic support task prompts: customer issue triage, account configuration checks, runbook lookup, and fix-versus-escalation decisions.
• Write scoring rubrics that define the correct routing decision, resolution path, or escalation criteria in terms that can be objectively checked.
• Set up task environments with ticket queues, account records, runbooks, and system states that replicate a genuine support scenario.
• Solve each task yourself to confirm the scenario is realistic, the correct answer is clear, and the rubric captures the right reasoning.
• Calibrate task difficulty — adjust ticket ambiguity, account configuration complexity, or runbook depth until the task challenges the model appropriately.
• Review and correct AI-drafted task prompts or rubrics when provided.

Qualifications

• Experienced support engineer, technical support specialist, or customer success engineer with hands‑on product support experience.
• Deep familiarity with real product failure modes, escalation paths, and support tooling (ticketing systems, internal knowledge bases, runbooks).
• Ability to reason through ambiguous support scenarios and articulate why a given routing decision is correct.
• Comfortable defining verifiable correctness criteria — not just "good customer service" but an objectively defensible resolution or escalation decision.
• Strong written communication; tasks and rubrics must be precise enough to be evaluated without interpretation.

Nice to Have

• Experience supporting enterprise or industrial software products.
• Background writing or maintaining runbooks, escalation playbooks, or internal knowledge bases.

Why Join Us

• Work on cutting‑edge AI projects alongside leading research labs.
• Fully remote and flexible — work when and where it suits you.
• Freelance autonomy with the structure of meaningful, task‑based work.
• Contribute directly to how AI agents handle real enterprise support scenarios.
• Potential for ongoing work and contract extension as new projects launch.