Project Duration: August 7 – December 31, 2025 (Remote)
Application Deadline: August 5, 2025
Email: [email protected]
Subject Line: AI BRIDGE MACHINE LANGUAGE EXPERT
Background
AfriLabs is the largest pan-African network of innovation and technology hubs, spanning 53 African countries and the Diaspora. With a mission to foster entrepreneurship and innovation, AfriLabs provides critical support through capacity-building programs, hackathons, incubation, and cross-border collaboration.
Secretariat Address: 18 Khartoum Street, Wuse Zone 5, Abuja, Nigeria
📧 Email: [email protected]
About the Partners
The Gates Foundation, through its Gender Equality Digital Connectivity (GEDC) and Digital Public Infrastructure (DPI) teams, is championing equitable digital access across Africa—especially for women.
Meta (formerly Facebook), a global AI leader, backs the Llama Impact Grants to encourage the development of inclusive, gender-sensitive AI tools using open-source LLMs like Llama.
Together, the Gates Foundation and Meta are supporting efforts to enhance AI outputs in African languages—making them more inclusive, accurate, and culturally relevant.
About the Project – AI BRIDGE
The AI Bias Reduction & Inclusive Data Generation Engine (AI BRIDGE) is a flagship program by AfriLabs designed to empower African innovators in developing AI-driven tools that are gender-sensitive, linguistically diverse, and inclusive.
- Phase 1: A hackathon involving 99 participants across 19 country-based teams.
- Phase 2: The top four teams now advance to develop a gender sanitization engine for 17 African languages with mentorship and expert support.
Project Objectives
- Detect and mitigate gender bias in training datasets for AI/LLM systems.
- Equip developers and curators with tools to rewrite biased content.
- Foster inclusive AI development across African languages.
- Support modular, scalable solutions that work across AI pipelines.
Project Methodology (Phase 2)
The project will follow a structured, rigorous, and culturally grounded methodology, including:
- Data Ingestion: Curating diverse African datasets aligned with gender inclusion.
- Bias Detection: Using hybrid scanning tools to flag problematic content.
- Content Analysis & Tagging: Classifying flagged entries with contextual metadata.
- Content Sanitization: Using transformer-based models for rewriting.
- Human-in-the-Loop Validation: Expert review by gender and linguistic specialists.
- Dataset Finalization: Traceable, version-controlled datasets aligned with ethical AI.
- Feedback Loops: Continuous improvement of classifiers and rewriting logic.
Purpose of the Assignment
AfriLabs seeks a Machine Language Expert to provide technical leadership to participating teams. The consultant will:
- Guide ML architecture design for gender bias detection and rewriting.
- Support development of inclusive NLP pipelines using transformer models.
- Advise on contextual generation strategies and feedback loops.
- Ensure alignment with ethical and linguistic standards across African languages.
Scope of Work
Key responsibilities include:
a) ML Architecture Advisory
- Recommend models and methods for bias detection and classification.
- Evaluate rule-based, statistical, and deep learning trade-offs.
b) Bias Detection Development
- Guide fine-tuning of models for gender bias identification using NLP techniques.
- Integrate explainability tools (e.g., SHAP, LIME).
c) Rewriting & Generation
- Recommend use of models like GPT, T5, mBART.
- Guide controlled text generation aligned with inclusive norms.
d) Feedback Integration
- Implement active learning and few-shot learning workflows.
- Guide version control and model updates using reviewer feedback.
e) Performance Evaluation
- Define model evaluation metrics (bias detection, fluency, semantic consistency).
- Support cross-validation and benchmarking strategies for low-resource languages.
f) Tooling & Infrastructure
- Recommend ML frameworks, experiment tracking (e.g., MLflow), and MLOps tools.
- Ensure ethical AI and data protection compliance.
g) Knowledge Transfer
- Conduct knowledge-sharing sessions on fairness-aware ML development.
- Mentor teams on responsible, explainable, and inclusive AI practices.
Required Qualifications
General
- Advanced degree in AI, machine learning, computational linguistics, or related field.
- Experience working with multilingual or under-resourced NLP datasets.
Specialized (any 2 or more):
- Expertise in designing fairness-aware NLP models (BERT, GPT, T5, etc.).
- Experience in text generation, paraphrasing, or rewriting tasks.
- Background in mentoring interdisciplinary AI teams.
- Ability to explain complex AI concepts in inclusive and ethical contexts.
Timeline
- Duration: 5 months
- Start Date: August 7, 2025
- End Date: December 31, 2025
- Mode: Remote
How to Apply
For Individuals
Submit the following to [email protected] (Subject: AI BRIDGE MACHINE LANGUAGE EXPERT):
- Filled Vendor Request Form (download link provided in the call)
- Cover Letter and CV/Portfolio
- Means of Identification
- Financial Proposal (in USD)
- 1–2 References (Name, Email, Phone)
- Sample of related work (preferably in innovation or grant programs)
For Companies
Submit the following:
- Filled Vendor Request Form
- Certificate of Incorporation
- Company Registration Document
- Cover Letter and CV of Team Lead
- Team Lead’s ID
- Financial Proposal (in USD)
- References and relevant sample work
Evaluation Criteria
- Technical Proposal – 65%
Proof of qualifications, experience, and relevance of submitted materials. - Financial Proposal – 35%
Proposal must be in USD and indicate if negotiable. Valid for at least 6 months.
Important Notes
- Only shortlisted applicants will be contacted.
- Women and Paid AfriLabs Hub Members are strongly encouraged to apply.
- Vendor Form: Download and use the attached file—do not modify the original template.
Secretariat Address: No 18, Khartoum Street, Wuse Zone 5, Abuja, Nigeria Email for Inquiries & Submissions: [email protected]
