Data Scientist

Responsibilities:

Analyze production logs to identify recurring failure patterns. Design and implement custom evaluations that go beyond standard metrics. Fine-tune and align models (LoRA, full fine-tuning, DPO) for domain-specific tasks. Build and maintain RAG pipelines with a focus on precision and recall. Create guardrail systems (classifiers, rule layers, confidence scoring) to filter or correct model outputs. Document failure modes and share findings across engineering and product teams. Partner with engineers to deploy models into production environments. Own the full lifecycle: data collection, model training, evaluation, deployment, monitoring.

Requirements:

Minimum of 4+ years’ experience in data science or machine learning engineering. 2+ years’ experience working with LLMs in production environments. Experience in Python Programming and a proven track record of building and deploying custom AI models (not just API wrappers). You must have created projects and solutions for LLMs to solve and they failed to solve and they failed to solve the problem accurately. Deep understanding of transformers, attention, and model architectures. Experience with PyTorch, HuggingFace Transformers, and LangChain or similar frameworks. Strong Knowledge of retrieval systems (vector databases, hybrid search, reranking). Familiarity with MLOps tools (MLflow, Weights and Biases, Kubernetes). Solid software engineering practices: testing, version control, code review.

Remuneration:  NGN 600,000