Snapshot 3723
Normalized text
Scripts and page chrome removed; this is what change detection compares.
Resources – Security & Compliance Enterprise Security, Fully Documented Vetted contributors working in controlled environments across the US, Thailand, and India. In-region storage, full audit trails, AES-256 at rest and TLS 1.2+ in transit. SOC 2 Type II attested ISO 27001 Certified ISMS In-region Data residency + audit logs Vetted Background checked workforce Security & Compliance Built for Enterprise Diligence Authentication, encryption, and governance documented to the standard your security team expects. Each control mapped to the framework, the policy, and the proof. Authentication MFA + Zero Trust Enterprise MFA with zero-trust access. Role-based permissions across teams, languages, and projects. Restricted internet, email, and ports on all annotation workstations. Encryption AES-256 + Key Control AES-256 encryption at rest, TLS 1.2+ in transit. Customer data isolated per project, without co-mingling across engagements. Audio, text, and metadata stay inside authorized environments. Data Governance Continuous Monitoring Real-time monitoring against GDPR and ISO 27001 controls. In-region storage, role-based access, and full audit logs — regulated data ready, end to end. PPH In-house tooling PPH In-house tooling pZero Controls are enforced in pZero, our proprietary annotation and evaluation platform. pZero provides role-based access, project-level data isolation, and audit logging across all engagements. Document Library Compliance & Assurance Documents Certifications ISO 27001 Certificate Download Document Full Reports SOC 2 Type II Report ISO 27001 Report Trust & Compliance Protect Data. Reduce Risk. Deliver with Confidence. ISO 27001:2022 Productive Playhouse meets the internationally recognized gold standard for information security, including 93 security controls across 14 domains. SOC 2 Type II Annual third-party audits ensure continuous compliance with all five trust criteria: Security, Availability, Processing Integrity, Confidentiality, and Privacy. Vendor Risk & Due Diligence FAQs What we evaluate, how we scale, and how we handle your data. All What We Evaluate How We Scale How We Handle Your Data All What We Evaluate How We Scale How We Handle Your Data All What kinds of problems do you actually solve? The ones that quietly kill model performance: LLMs that sound right but are wrong, search that misses intent, outputs that break outside English, edge cases nothing else catches. What exactly do you evaluate? LLM responses for accuracy and reasoning. Search for relevance and ranking. Personalization for fit. News and UGC for credibility. Media for quality. Golden datasets to validate the rest. How do you handle multilingual complexity? Native-level raters who understand context, culture, and intent — not just translation. The difference between a model that works globally and one that quietly fails outside English. How do you scale without breaking quality? Work is structured as measurable tasks, assigned to trained raters, tracked to time per task. Predictable throughput, fast capacity flex, consistent quality. No chaos, no black box. What happens as volume and complexity increase? More volume and harder tasks mean more human effort. What matters is how it’s managed. We align task design, effort, and workforce so you scale without degrading quality or timelines Can our data be used to train your models or third-party models? No. Customer data is used only to deliver the contracted service. We do not train internal or third-party models on your data without explicit written authorization. Which certifications and frameworks do you maintain? SOC 2 Type II and ISO 27001, with continuous monitoring against GDPR, CCPA, and HIPAA controls. Reports available under NDA via the document library. Do you isolate customer data across engagements? Yes. Each engagement runs in a dedicated environment with no co-mingling of customer data, models, or annotations across projects or clients. What controls exist on annotation workstations? Role-based access, MFA, restricted internet and email, blocked ports, and disabled local printing. Annotation happens inside our proprietary tooling, not in shared environments. What We Evaluate What kinds of problems do you actually solve? The ones that quietly kill model performance: LLMs that sound right but are wrong, search that misses intent, outputs that break outside English, edge cases nothing else catches. What exactly do you evaluate? LLM responses for accuracy and reasoning. Search for relevance and ranking. Personalization for fit. News and UGC for credibility. Media for quality. Golden datasets to validate the rest. How do you handle multilingual complexity? Native-level raters who understand context, culture, and intent — not just translation. The difference between a model that works globally and one that quietly fails outside English. How We Scale How do you scale without breaking quality? Work is structured as measurable tasks, assigned to trained raters, tracked to time per task. Predictable throughput, fast capacity flex, consistent quality. No chaos, no black box. What happens as volume and complexity increase? More volume and harder tasks mean more human effort. What matters is how it’s managed. We align task design, effort, and workforce so you scale without degrading quality or timelines Can our data be used to train your models or third-party models? No. Customer data is used only to deliver the contracted service. We do not train internal or third-party models on your data without explicit written authorization. How We Handle Your Data Which certifications and frameworks do you maintain? SOC 2 Type II and ISO 27001, with continuous monitoring against GDPR, CCPA, and HIPAA controls. Reports available under NDA via the document library. Do you isolate customer data across engagements? Yes. Each engagement runs in a dedicated environment with no co-mingling of customer data, models, or annotations across projects or clients. What controls exist on annotation workstations? Role-based access, MFA, restricted internet and email, blocked ports, and disabled local printing. Annotation happens inside our proprietary tooling, not in shared environments. Share Your Specs Tell us your security requirements and we’ll map our controls to your framework, your policies, and your proof. Contact Us