Human data and evaluation for AI labs.

Human data is the work people do to train and evaluate AI models: preference rankings, rubric grading, written reference answers, red-team prompts and agent trajectory reviews. Awana provides it with calibrated software engineers who understand the outputs they grade.

Awana provides preference and RLHF data, rubric-based evaluation, agent and computer-use trajectory grading, red-team sprints, golden sets and rubrics, and multilingual evaluation in English, Spanish and Portuguese.

Awana's evaluators are software engineers, hundreds of whom grade AI output for clients today. Awana is United States-led, with a cost-effective Latin America engineering team that works in US time zones.

Yes. Awana evaluators grade multi-step trajectories: tool choice, arguments, plans, intermediate states and whether the task actually got done, using your rubric and your tooling.

Every Awana evaluator passes a domain screen, a skills test on a real task, and a blind calibration test against your gold set before touching live data. When your rubric changes, Awana re-calibrates the pod before the next batch.

Awana tracks inter-rater agreement on every batch against a floor you set, runs second-pass review on a sample, and logs ambiguous items back to your researchers so the rubric gets sharper over time.

Yes. Awana runs red-team sprints under your safety policy and access rules, writing adversarial prompts and attack sets and documenting each successful attack for your team. Awana does not run bio, chemical or nuclear risk testing.

Yes. Awana recommends a short, paid pilot on one task with a small calibrated pod, so you can judge quality, agreement and throughput before committing to a larger pod.

Yes. Awana evaluators work in your labeling platform, internal tools or virtual desktops, through your SSO, with least-privilege access that is removed the day someone rolls off.

Awana signs an NDA and standard DPA before any access, background-checks every evaluator, and keeps all work inside your environment. Awana works under your access tiers and redaction rules.

Yes. Awana evaluators can work under tiered access, where your team controls what each tier sees and what can be shared, with logged sessions and your redaction rules applied.

Yes. Awana is United States-led, and its leadership holds an active US security clearance. For public-sector work, Awana scopes staffing to the program's requirements, including US-based personnel where required.

Awana typically has first evaluators in days and most pods staffed in about three weeks, including calibration against your gold set.

Awana keeps a dedicated core on your project and adds calibrated evaluators for launches or eval pushes, then scales back, month to month.

Awana works hourly or as a fixed-price project. Pilots are fixed-price projects; dedicated pods and surge capacity are billed hourly per evaluator. Awana gives you a quote after the first call.

Awana provides calibrated software engineers for human data and AI evaluation, from preference data to agent trajectory grading and red-team sprints. Awana is United States-led, works in US time zones, and works hourly or as a fixed-price project.