Consultant Role | Content Expert - AI Evaluation, Agricultural Content, Research & MEL |
Engagement Period | Six months, with possibility of extension |
Location | Nigeria / Remote, with field participation as specifically authorized |
About Digital Green
Digital Green is a global non-profit that has spent nearly two decades helping smallholder farmers build more prosperous, resilient livelihoods—reaching over 10 million farmers to date. Our flagship product, FarmerChat, is a multilingual AI-powered advisory platform that gives farmers timely, locally relevant guidance—from climate-resilient practices to pest management, input use, and market access—in the language and format that works for them. Today FarmerChat reaches farmers across India, Kenya, Ethiopia, Nigeria, and Brazil, with millions of questions asked and answered in over 15 local languages. We combine deep, on-the-ground agricultural extension experience with modern AI to meet farmers where they are, and help them make better decisions, faster.
1. Background
Role Summary
We are seeking an experienced Nigeria-based agricultural content expert who can support AI response validation, agricultural content and terminology review, and research, product testing, and field/program activities. The ideal candidate combines strong local crop and/or livestock knowledge with the ability to review information carefully, communicate clear recommendations, and collaborate with remote and field teams. Assignments across the three workstreams will vary based on program needs.
Digital Green is enhancing its AI-driven agricultural advisory systems, including FarmerChat, by improving the agronomic accuracy, contextual relevance, localized applicability, usability, and safety of AI-generated responses and farmer-facing agricultural content.
The Consultant will serve as an independent Content Expert with the expertise to support all three core workstreams:
- Workstream A: Reviewing and Checking the Quality of AI-Generated Agricultural Advice
- Workstream B: Developing and Reviewing Agricultural Content, Crop Calendars, and Local Terminology
- Workstream C: Ad hoc support—Research, Monitoring, Evaluation & Learning (MEL), Product Testing, and Program Support
Actual assignments will depend on the Consultant's capabilities and Digital Green's needs. All assignments must be agreed upon in writing with Digital Green's designated Technical Lead before work begins. The scope, volume, timing, and cadence of work will be determined for each assignment. Work under any particular workstream is not guaranteed.
2. Workstream A - Reviewing and Checking the Quality of AI-Generated Agricultural Advice
The Consultant shall perform independent quality validation of agricultural Q&A pairs and AI-generated outputs that have undergone first-level expert review.
- Independently review outputs assessed by Content Expert Reviewers.
- Verify whether reviewer assessments and corrections meet Digital Green's quality standards for agronomic and technical accuracy, relevance, completeness, clarity, local agricultural context, farmer usability, safety, and appropriateness.
- Validate assigned evaluation formats, including text, image, MCQ, and other formats specifically assigned by Digital Green.
- Confirm, correct, or reject reviewer assessments where appropriate.
- Identify inconsistencies, reviewer errors, model hallucinations, recurring quality issues, and systemic patterns.
- Provide corrective feedback to improve reviewer consistency and overall evaluation quality.
- Support the development and maintenance of reliable Golden Q&A datasets, evaluation benchmarks, and model-quality datasets.
- Record and submit final validation decisions through Digital Green's designated platform, spreadsheet, dashboard, or other approved workflow.
3. Workstream B - Developing and Reviewing Agricultural Content, Crop Calendars, and Local Terminology
Upon written assignment from Digital Green, the Service Provider may:
- Perform technical, agronomic, and localized review of agricultural Content Cards.
- Review AI-generated or manually drafted agricultural advisory content and provide technical corrections, edits, recommendations, and supporting references or citations where required.
- Develop, adapt, update, or review crop calendars and seasonal advisory workflows.
- Tailor agricultural content to relevant crops, livestock systems, agro-ecological zones, local seasons, farming systems, and agricultural practices.
- Develop, review, validate, and map locally relevant agricultural terminology, including terminology related to crops, livestock, pests and diseases, agricultural inputs, agronomic practices, measurements, and other domain-specific concepts.
- Support development and improvement of agricultural glossaries, terminology datasets, taxonomies, thesauri, and structured agricultural knowledge resources.
- Undertake other agricultural content and knowledge activities as specifically authorized.
4. Workstream C - Supporting Agricultural Research, Farmer Feedback, Field Activities, and Product Testing
- Review agricultural research reports, data, and farmer feedback to ensure the information is accurate and relevant to local farming practices.
- Support monitoring and evaluation activities, including reviewing farmer surveys, field observations, and program results.
- Provide technical input for agricultural research, field studies, and program activities.
- Test FarmerChat and its new features to check the accuracy, usefulness, and clarity of agricultural advice.
- Identify issues in agricultural content or product performance and suggest improvements.
- Support field activities, pilot programs, and farmer feedback sessions, including coordination, documentation, and sharing observations.
- Participate in partner discussions, technical workshops, training sessions, and program meetings as required.
- Prepare brief reports, summaries, and recommendations based on research findings, field activities, and program needs.
Participation in technical workshops, training activities, stakeholder working groups, or other external engagements must be assigned and approved in writing by Digital Green in advance. The Consultant may not represent or speak on behalf of Digital Green without express written authorization and must immediately notify Digital Green of any external requests concerning FarmerChat or other Digital Green products or activities before accepting or responding substantively.
5. Work Authorization and No Minimum Commitment
The Consultant shall perform Services only as expressly authorized by Digital Green through this SOW, a written task order, assignment, email instruction, or other written authorization from Digital Green's designated Technical Lead. Any estimated volumes, hours, workload, frequency, or timing are non-binding planning estimates only and do not create a guaranteed minimum commitment.
The Consultant shall not exceed authorized work or level of effort, incur expenses, or undertake additional activities outside the approved scope without prior written approval from Digital Green.
6. Quality Review and Acceptance
Digital Green may return deliverables for correction if they are incomplete, inaccurate, technically incorrect, inconsistent with applicable guidelines, or otherwise fail to meet required quality standards. Reasonable corrections associated with the original assignment shall be completed as part of that assignment.
7. Progress Updates and Communication
- Participate in alignment and calibration meetings as reasonably requested.
- Maintain accurate records of completed units and hours.
- Provide progress updates when requested by the Technical Lead.
- Submit deliverables through Digital Green's designated platform, spreadsheet, dashboard, or other approved workflow.
8. Minimum Qualifications
Education: Advanced degree in Agriculture, Agronomy, Agricultural Extension, Veterinary Sciences, Horticulture, or a related discipline, or equivalent demonstrated professional expertise.
Experience: Minimum of five years of relevant experience in agricultural advisory services, extension, research, veterinary/agronomic services, agricultural content development, or related technical work.
Technical Expertise: Strong understanding of smallholder farming systems, localized agricultural practices, crop and/or livestock management, and farmer advisory workflows.
Preferred: Experience with digital agriculture, AI-assisted advisory systems, human-in-the-loop AI evaluation, RLHF, agricultural content development, quality validation, or digital product testing.