Candidates will primarily work on domain-specific tasks such as:
Fine-tuning:
Curating datasets of prompts/questions and responses that are accurate, comprehensive, and contextually appropriate.
RLHF / Human Evaluation:
Assessing model outputs in Physics and Mathematics domains relative to user prompts/questions.
Example Tasks:
Developing undergraduate-level
Physics questions and answers
for ChatGPT prompts.
Constructing
Mathematics benchmarks
and evaluating model performance with precise answers.
(Secondary: Supporting Chemistry and Biology-related tasks if required.)
2. Time Commitment & Availability
Weekly Hours:
Flexible (10-40 hrs/week), with ~20 hrs/week as standard.
Overlap Requirement:
At least 4 hrs/day PST overlap.
3. Project Duration & Timeline
Typical engagement:
1 month
, with extension possible based on performance and project pipeline.
4. Compensation & Payments
Range:
$35-$70/hour (based on expertise and domain).
Payment Mode:
Hourly, invoiced.
5. Hiring & Evaluation Process
Submissions via the
Agency Portal
+ Interest Check Form.
Process may include:
Direct onboarding (no interview)
Domain-focused interview
Short evaluation task
6. Visa & Work Authorization
No visa sponsorship available.
Candidates must be eligible to work independently in the specified geography.
Offer Snapshot
Commitment:
20-40 hrs/week, with PST overlap.
Type:
Contractor assignment (no paid leave/medical benefits).
United States, Canada, Western Europe (UK, France, Germany, Switzerland, Denmark, Finland, Netherlands, Sweden, Iceland, Italy, Austria, Ireland, Norway)