AI Research Lead - African First Language Models, London

Department, Team, Location & Position

Department: Technology

Team: AI Foundry

Location: London (Hybrid or Remote) / Nairobi

Position type: Contractor (umbrella – inside IR35) until 24th December 2026; Full Time

What The Hiring Manager Says

\"The Africa AI Language Initiative is one of our most strategically important efforts — aiming to build Africa-first, open, locally representative AI models across the continent’s most widely spoken languages. As our Subject Matter Expert, you will play a pivotal role in shaping model design, guiding technical execution, and ensuring high-quality delivery across multiple partners.\" — Director of AI Initiatives

About The Team

The Africa AI Language Initiative brings together mobile operators, African researchers, startups, academia, and GSMA to build and deploy open, representative language models for Africa. The team is multidisciplinary — spanning AI research, data governance, telco engagement, technical programme delivery, and community development.

About The Role

You will guide the development of 3–5 Africa-first open language models end-to-end, acting as the technical anchor and coordinator across multiple model‑building tracks. You will ensure alignment between researchers, operators, and partners, while driving high-quality technical outcomes, documentation, and model readiness.

About You

You will bring demonstrable experience in developing and evaluating language models, with a strong grounding in NLP techniques relevant to African languages. You’ll be able to work across diverse linguistic structures—such as tonal variation, agglutination, rich morphology, and non-standardised orthographies—to ensure that models reflect real-world language use across the continent. Your ability to analyse datasets, identify representational gaps, and guide the creation of high-quality corpora will directly shape the performance and inclusiveness of our Africa first models.

You’ll be comfortable working end-to-end across the research workflow: from tokenisation strategies suitable for low resource languages, to training and finetuning multilingual models, to designing evaluation methods that capture linguistic nuance. You will apply your technical depth to help partners understand model behaviour, tradeoffs, and opportunities for improvement. You’ll communicate these insights clearly, helping a broad ecosystem—including operators, researchers, community groups, and developers—collaborate effectively around open model development.

You’ll be curious and adaptable, able to test new methods quickly while maintaining scientific rigour. Through structured documentation, reproducible experimentation, and responsible data practices, you’ll help ensure these open models meet the highest standards for transparency and reliability.

Core Skills

• Data Analysis: apply analytical rigour to evaluate datasets and model outputs, identifying linguistic patterns, bias risks, and areas for improving representation across African languages.
• AI Research: explore and adapt advanced NLP techniques—including tokenisation, alignment, and evaluation methods—to address the challenges of low-resource and morphologically rich languages.
• Communication: translate complex technical concepts into clear, accessible explanations for partners across Africa, enabling shared understanding of modelling decisions and outcomes.
• Delivery: drive research workstreams with structure and consistency, ensuring training runs, evaluations, documentation, and model releases are delivered to a high technical standard.
• Stakeholder Management: collaborate confidently with researchers, operators, universities, and data communities, ensuring alignment on linguistic priorities, data quality, and model behaviour.

Contract type

Short term Contractor

What We Offer

In addition to architect-designed offices and competitive compensation, our benefits include fantastic learning & development opportunities, generous holiday allowances, four additional days off for professional development and many others.

#J-18808-Ljbffr