AI Domain Architect

About the role

Softgic is looking for an experienced AI Domain Architect in Nairobi to design and build the domain intelligence layer behind enterprise AI solutions. This is a senior individual contributor, hands-on role combining architectural thinking with practical AI engineering skills, taking ideas from concept to working Proof of Concept and then to scalable enterprise standards.

Key responsibilities

• Design and evolve ontologies, knowledge graphs, semantic layers and canonical domain models for enterprise platforms, defining entities, relationships, business rules and terminology.
• Define agentic AI workflows, including task decomposition, tool boundaries, escalation paths and human-in-the-loop checkpoints.
• Establish when to use RAG, fine-tuning or prompting, and set standards for data preparation, labeling, provenance, governance and retention.
• Build domain-specific model evaluation pipelines and reusable AI patterns such as retrieval strategies, reasoning templates and agent scaffolds.
• Build PoCs to validate architectures and patterns before making them standards.
• Work hands-on with Microsoft Foundry and Azure for model deployment, agent development, evaluation and AI safety.
• Partner with product and engineering teams from design through go-live, create technical standards and documentation, and act as a technical escalation point.

Required profile

• 8+ years of professional software engineering experience in production environments, at Staff Engineer, Principal Engineer, Architect or equivalent senior level.
• Hands-on experience designing ontologies, knowledge graphs or semantic layers deployed in production.
• Practical experience with LLMs and agentic AI, and with RAG, fine-tuning and model evaluation pipelines.
• Ability to move an AI pattern from concept to PoC to published standard, and to influence technical direction across teams.
• Excellent written and verbal communication skills and advanced English; a plus is experience in regulated domains such as PEO, HCM or payroll.

Required skills

• Microsoft Foundry and Azure (compute, data, identity, networking).
• Nice to have: OWL, RDF, SHACL, property graphs, multi-tenant SaaS, SOC 2 or ISO 27001, Microsoft Agent Framework, MCP, and Snowflake Cortex AI.