The Consultant will work under the direct supervision of the Head of Experimentation, Accelerator Lab, UNDP to undertake the following tasks:
Coordinate with Technology Partners
• Coordinate with Technology Partners for development of digital public good and work in close coordination with all stakeholders, translate the high-level requirements into use-cases and technical requirements.
• Coordinate and manage data partnerships through data sharing agreements and data collaborative strategy as required by the digital public good.
• Provide timely inputs for developing systems architecture, data pipeline, infrastructure requirements, and aspects of data quality
• Ensure adherence to UN Principles on Personal data protection and privacy, data security, data confidentiality and access.
• Ensure robust prediction model considering aspects model accuracy, bias-variance tradeoff, fairness, accountability, transparency, ethics.
• Provide inputs to customize reports and analytics for different user-categories as per the requirement of the project.
Peer Review Algorithms
• Facilitate peer review process of DiCRA analytics and openAI algorithms. Establish multilevel peer review process, document the guidelines and publish them transparently on Github
• Identify and onboard atleast 10 qualified peer reviewers with specialization in Geospatial data science for reviewing DiCRA algorithms
• Ensure publication of peer-review documents for each analytics layer under DiCRA on Github; and incorporation of feedback on DiCRA platform
DiCRA collaborative
• Facilitate, foster and expand the DiCRA collaborative through structured processes, data coordination and data pipeline mechanisms.
• Facilitate atleast 20 meetings of DiCRA data collaborative and ensure continuous publication of new data layers and use cases.
• Establish mechanisms for sharing cloud computing credits with contributing data scientists.
• Develop and publish openAI algorithms for complex use cases such as crop diversity, farm boundaries, oil palm, polyhouse detection etc.
User-Feedback and Functionality improvement
• DiCRA is developed for the agriculture ecosystem with diverse stakeholders who have diverse requirements.
• Conduct user feedback sessions and collect feedback from diverse users and improve overall functionality of DiCRA
• Improve the Use-case section under DiCRA through automated processes
AI strategy
• Provide inputs on proposals and concept notes for other innovative applications of AI in domains such as climate, air pollution, energy, natural resource management, skills etc., as and when required by the country office.
Expected Deliverables:
|
S.No |
Deliverables |
Amount |
|
1 |
Establish peer review guidelines for DiCRA algorithms and upload them on Github Identify and onboard atleast 10 qualified peer reviewers for review of DiCRA algorithms |
15% |
|
2 |
Publish peer-review documents for each analytics layer under DiCRA on Github Report on incorporation of feedback in DiCRA from peer-review |
20% |
|
3 |
Collect feedback from diverse users and improve overall functionality of DiCRA Improve the Use-case section under DiCRA through automated processes |
15% |
|
4 |
Establish mechanisms for sharing cloud computing credits with contributing data scientists Facilitate atleast 20 meetings of DiCRA data collaborative and ensure continuous publication of new data layers |
20% |
|
5 |
Develop and publish complex Use cases on crop diversity, farm boundaries, oil palm, and polyhouses |
30% |