NLP Engineer — Multilingual (French / Arabic / Swahili)

Role overview

A specialist NLP role focused on building and improving sentiment and policy-signal extraction natively in French, Arabic, and Swahili. The mandate is to read the languages in which African investment decisions are actually made, preserving nuance that English-pivot pipelines routinely lose. The work feeds directly into the platform's core market and sentiment signals as a first-class input rather than a derived feature.

Responsibilities

- Build and improve sentiment and policy-signal extraction natively in French, Arabic, and Swahili. - Process source languages directly rather than translating to English first, keeping nuance intact. - Expand regional source coverage and adapt models to region-specific media registers. - Feed the resulting output into the core market and sentiment signals as a first-class input. - Document where translation pipelines quietly destroy policy nuance and design the system around those failure modes.

Requirements

- Multilingual NLP experience, ideally in lower-resource language settings. - Working fluency in at least one of French, Arabic, or Swahili. - Clear-eyed understanding of the failure modes where translation pipelines strip out policy nuance. - Comfort working with region-specific media registers and uneven source quality.

Nice to have

- Experience adapting transformer-based models to low-resource African languages. - Background in policy-text analysis, sentiment analysis, or both.