HelpMum
Healthcare
AI Engineer
Share this role
About this role
HelpMum—an award-winning, venture-backed social enterprise leveraging mobile technology and artificial intelligence to eliminate maternal and infant mortality across Africa—is recruiting an AI Engineer for its healthcare technology lab in Lekki Phase 2, Lagos State. Reporting to the Head of AI & Engineering, this role designs, fine-tunes, and deploys production-grade Large Language Models (LLMs), Natural Language Processing (NLP) frameworks, Speech AI tools (STT/TTS in local African languages), and scalable MLOps pipelines optimized for low-bandwidth healthcare environments.
Key Responsibilities
-
LLM & Generative AI Development: Architect, fine-tune, and deploy domain-specific Large Language Models (LLMs) and NLP pipelines for maternal/infant medical guidance and automated healthcare triage.
-
Speech AI & Local Language Modeling: Build Speech-to-Text (STT) and Text-to-Speech (TTS) conversational AI engines tailored for local African languages and accents.
-
MLOps & Low-Bandwidth Infrastructure: Deploy scalable ML model pipelines using Docker, Kubernetes, and cloud infrastructure (AWS/Azure/GCP) designed to operate reliably in low-connectivity settings.
-
Dataset Curation & Model Validation: Clean, curate, and preprocess healthcare datasets for supervised fine-tuning, retrieval-augmented generation (RAG), and responsible AI evaluation.
-
Academic Credentials: Bachelor's or Master's Degree in CS, Data Science, AI, or technical equivalents.
-
Core Tech Stack:
-
Languages & Frameworks: Expert Python, PyTorch, TensorFlow, Keras.
-
NLP & GenAI: Transformers, Hugging Face, LLM Fine-Tuning, Speech AI (STT/TTS), RAG architecture.
-
Data & Cloud: Advanced SQL, Data Pipelines, AWS/Azure/GCP.
-
MLOps & DevOps: Docker, Kubernetes, Model Monitoring, CI/CD pipelines.
-
-
Core Competencies: Speech AI modeling, African language translation models, MLOps deployment, data engineering, and resource-constrained software optimization.
This role accepts applications on an external page.