This role is for one of our clients
Compensation: $70 - $100 per hour
We're seeking advanced Excel experts with deep, hands-on experience in finance — professionals who live in spreadsheets, building financial models and analyses at a senior, professional level. Your expertise will directly elevate the quality of the training data behind the most advanced Large Language Models.
This role is for finance professionals who apply advanced Excel — three-statement modeling, valuation, forecasting, and data analysis — in domains such as private equity, investment banking, hedge funds, and equity research (accounting / FP&A also welcome).
Requirements
Key Responsibilities
- Guide research teams to close knowledge gaps and improve AI model performance on Excel-based tasks within finance.
- Design realistic, domain-relevant Excel tasks drawn from your professional experience, and write accurate, well-structured solutions.
- Evaluate Excel tasks and solutions and provide clear, written feedback.
- Collaborate with other subject matter experts to ensure consistency and accuracy in training data.
Core Qualifications
- Demonstrated advanced, professional-level Excel expertise, including advanced formulas (e.g., XLOOKUP, INDEX/MATCH, dynamic arrays), PivotTables, financial modeling, dashboards, Power Query, Power Pivot, and macros/VBA.
- Hands-on professional experience applying Excel within a finance domain — private equity, investment banking, hedge funds, or equity research (accounting / FP&A also considered), including work such as three-statement models, LBO/DCF and valuation, and forecasting.
- Demonstrable career progression in your field.
- Prior experience in AI training, human data annotation, or model evaluation — ideally Rubrics-trained on Mercor or a comparable platform — is a strong plus.
- Ability to engage reliably for 40 hours/week during weekdays through the end of September, then 20 hours/week thereafter.
- Verbal and written communication skills, problem-solving skills, and interpersonal skills.
- United States–based only.