Responsibilities
- Monitor inbound product mentions on a continuous sweep cadence (queue polled roughly every 2 minutes) and apply a first-pass tier within the 30-minute live-pass latency target.
- Make the elevate-or-not call with high precision.
- Log Tier 3 posts to the signal store with no outreach.
- Own Tier 2 end to end. Respond to the user on-platform, ask the clarifying questions needed to establish a repro (prompt, settings, model, steps, screenshots), and drive the thread to a stated resolution.
- Extract and log the repro, assign an issue category, file or link the corresponding bug, and track it to closure so the user gets a real answer rather than an acknowledgement.
- Record tier, tier_reason, and scoring_mode on every item so misclassifications are auditable, and flag posts that are below threshold but trending for continued re-scoring.
- Draft and reuse approved response language for recurring product issues; elevate to comms when a situation needs net-new external language rather than improvising it.
- Hand off cleanly on shift change with a live queue state, including open threads, pending repros, and anything held for the maturity pass.
- Surface recurring themes from Tier 2 and Tier 3 volume to product and comms partners on a regular cadence.
Minimum Viable Product Expertise
Minimum viable product expertise (developed in ramp) Working fluency in product's surfaces, generation flows, model behavior, and known limitations, sufficient to tell a real defect from expected behavior, a user error, or a feature request. Familiarity with current known issues, shipped fixes, and public positioning, so the responder does not confirm something already resolved or speculate on anything unannounced. Ability to judge whether a complaint is concrete enough that an engineer could act on it, which is the actual Tier 2 bar.
Requirements
- Experience responding to users in public social channels on behalf of a brand or product, with a track record of de-escalating rather than inflaming.
- Demonstrated judgment under time pressure and incomplete information, with a bias toward escalating ambiguous reputational risk rather than absorbing it.
- Strong, concise written voice that holds up in public reply threads.
- Comfort with technical product detail: can read a stack of screenshots and prompt settings and turn it into a usable repro.
- Ability to work a shift-based schedule, including evenings and weekends, to sustain continuous low-latency coverage.
- Rigorous logging discipline. Every call is recorded with the rule that fired it.
Preferred Qualifications
- Prior work in a trust and safety, community, social care, or incident-adjacent triage function.
- Prior exposure to generative AI image or video products and the common failure modes users report.
- Familiarity with social listening and queue tooling.
Pay
- Pay range is $35 - $40 per hour with full benefits available, including paid time off, medical/dental/vision/life insurance, 401K, parental leave, and more.
- Our compensation reflects the cost of labor across several US geographic markets.
- Pay is based on several factors including market location and may vary depending on job-related knowledge, skills, and experience.
Our Promises
At Crystal Equation, we empower people and advance technology initiatives by building trust.
For more information regarding our Privacy Policy, please visit crystalequation.com/privacy.
#J-18808-Ljbffr