Hello, We are launching a language technology project for Chimini, a low-resource Bantu language, and are seeking an ML/NLP engineer to help us design and implement the foundational phase of the project. Long-Term Vision Our long-term goal is to build A structured Chimini text + audio corpus A scalable API layer for integration into our own applications Eventually, speech-to-text and text-to-speech capability in Chimini Chimini is historically related to Swahili, but we do not yet know how structurally similar they are. Pronunciation may differ significantly, which may impact model transfer for speech systems. We currently have Written texts Audio recordings Access to native speakers for transcription and validation Phase 1 (3–6 Months) The objective of Phase 1 is to build a strong ML-ready foundation, including Designing a scalable database structure for text and audio Preparing and structuring data for NLP workflows Building a clean corpus pipeline (segmentation, transcription storage, metadata) Advising on whether Chimini–Swahili linguistic comparison should be conducted before leveraging transfer learning Evaluating potential approaches Fine-tuning multilingual models Embedding-based retrieval systems LLM + RAG architectures Longer-term speech model strategy We want the system designed from the beginning to support future ML training and experimentation.

Responsibilities

Define ML/NLP strategy for a low-resource language Recommend architecture for scalable corpus and training workflows Implement foundational data pipelines Advise on transfer learning feasibility from Swahili or multilingual models Provide phased roadmap (short-term vs long-term) Ideal Experience NLP for low-resource or multilingual languages Speech systems (ASR/TTS) Fine-tuning transformer models Embeddings and vector databases Designing ML pipelines for scalable experimentation We will handle data collection, transcription, and language validation. Please include Relevant ML/NLP experience Proposed high-level technical approach Estimated timeline for Phase 1 Availability We are looking for someone who can help architect this correctly from the start, with long-term ML scalability in mind. Best regards, Apply To this Job Apply To this Job Apply To this Job