AI Solutions Architect

Location: Addis Abeba

Department: Technology, Group Product and Technology Office

Reports To: Group Chief Product and Technology Officer

About Kifiya:

Kifiya is an AI-powered financial and market infrastructure company advancing inclusive economic growth across Africa. We design and deploy risk decisioning systems, intelligent financial infrastructure, and market linkage platforms that unlock credit, insurance, payments, and capital access for MSMEs and smallholder farmers. We build the AI, data, and financial infrastructure that helps financial institutions serve hard-to-finance segments at scale.

Business Unit

Description The Group Product and Technology organization is Kifiya's enterprise technology backbone. Under the Group Chief Product and Technology Officer, it brings together the centralized Product Office, Intelligent Financial Infrastructure, Intelligent Data Decisioning, Customer Support Excellence, centralized Infrastructure and DevOps, and group-office staff functions as one integrated product and technology organization.

The organization designs, builds, governs, and operates the technology platforms that power Kifiya's AI-enabled financial, agricultural, capital, and market infrastructure solutions. It operationalizes Kifiya's infrastructure model across data, intelligent decisioning, financial rails, vertical solutions, and capital and risk infrastructure, ensuring that these capabilities operate as a secure, interoperable, scalable, and reusable ecosystem.

The Group Product and Technology Office establishes the common architecture, delivery, service, security, and technology standards that allow the separate delivery and service lanes to operate as one coherent organization. Its staff functions carry functional authority across the group while leaving delivery ownership, people management, backlog ownership, and operational accountability within the relevant lanes.

Position Summary

The AI Solutions Architect is Kifiya's company-wide functional authority for solution architecture and the adoption of AI by design.

The role maintains Kifiya's target and transition architecture, reference architectures, reusable solution and integration patterns, and Architecture Decision Record discipline. It owns the AI-by-design standard governing where and how artificial intelligence may be used, what data may be used, and what controls, human oversight, documentation, evaluation, security, and monitoring are required.

The AI Solutions Architect chairs the Architecture Review Board, determines whether material solution designs conform to approved architecture and AI-by-design standards, and may withhold architecture sign-off where a design does not conform. Where a valid business need requires deviation, the role prepares a documented exception recommendation for decision by the Group CPTO rather than allowing non-conformance to become an indefinite delivery block.

The role operates horizontally across the Product Office, Intelligent Financial Infrastructure, Intelligent Data Decisioning, Customer Support Excellence, Infrastructure and DevOps, CISO, and relevant corporate and business functions. It exercises influence through standards, evidence, design review, facilitation, and technical credibility rather than through line authority.

The role does not own product vision, commercial scope, backlog sequencing, engineering or DevOps capacity, application delivery, production operations, model development, or model performance. These accountabilities remain with the designated Product, Technical Product Management, IFI, IDD, Customer Support Excellence, Infrastructure and DevOps, and CISO leaders.

Key Responsibilities

1. Target and Transition Architecture

• Maintain Kifiya's current-state, target, and transition architecture in line with the product and technology strategy established by the Group CPTO.
• Translate Kifiya's strategic direction into clear architectural principles, capability maps, domain boundaries, target-state designs, and phased transition pathways.
• Define and maintain reference architectures and reusable patterns for:
• AI-enabled credit, risk, fraud, and decisioning systems.
• Intelligent Financial Infrastructure.
• Data and AI platforms.
• Financial and market linkage platforms.
• Multi-tenant products and partner-bank integrations.
• API, event, and data integration.
• Cloud, hybrid, and on-premise deployment.
• Multi-country and configurable regulatory deployment.
• Generative and agentic AI solutions, where applicable.
• Establish architectural guardrails covering modularity, interoperability, reuse, security, privacy, resilience, observability, scalability, performance, maintainability, cost efficiency, and responsible technology use.
• Promote API-first, standards-based, and reusable designs that reduce duplication and enable shared capabilities to be built once and consumed across products and business lines.
• Ensure architecture decisions reflect the operating, regulatory, customer, connectivity, and infrastructure conditions of Kifiya's target markets.
• Identify architectural dependencies, obsolescence risks, capability gaps, and material architecture-related technical debt.
• Recommend transition and remediation priorities and ensure required delivery work enters the relevant IFI or IDD backlog through the approved intake process.
• Maintain a clear distinction between setting architectural direction and owning delivery execution or backlog priority.

AI-by-Design and Responsible AI Architecture

• Own and continuously improve Kifiya's AI-by-design standard across traditional machine learning, intelligent decisioning, generative AI, and agentic AI use cases.
• Establish clear criteria for:
• When AI is appropriate and when deterministic or rules-based approaches are preferable.
• The level of risk and control required for different AI use cases.
• Permitted data sources and data-use conditions.
• Human review, intervention, override, and escalation.
• Explainability and transparency.
• Fairness, bias assessment, and inclusion.
• Privacy, confidentiality, and data minimization.
• Model and data provenance.
• Testing, evaluation, validation, and acceptance.
• Production monitoring, fallback, rollback, and retirement.
• Documentation, traceability, and audit evidence.
• Require an AI-by-design assessment during discovery for every initiative involving, or materially affected by, an AI component.
• Determine the architecture and AI control conditions that must be satisfied before an initiative is considered qualified for delivery.
• Establish reusable AI solution patterns that enable responsible adoption without requiring each team to design governance and controls from the beginning.
• Define additional architecture controls for generative and agentic AI, including controls for:
• Input and prompt manipulation.
• Sensitive-information exposure.
• Model, data, and software supply-chain risk.
• Unsafe or excessive tool access.
• Inadequate output validation.
• Identity and privilege management.
• Uncontrolled consumption and cost.
• Human confirmation before high-impact actions.
• Ensure AI solutions include appropriate evaluation, observability, audit trails, human escalation, and post-deployment monitoring from the point of design.
• Partner with the Chief Data Officer, Head of Data Science, CISO, Product Office, Infrastructure and DevOps, and relevant Risk, Compliance, Legal, and Sharia authorities to translate governance requirements into implementable architecture conditions.
• Provide architecture direction without taking ownership of model development, model validation, data fitness decisions, production model performance, or model retraining.

Architecture Governance and Decision Management

• Chair the Architecture Review Board and ensure it operates as a decision-making forum rather than a documentation or presentation forum.
• Establish a risk- and materiality-based architecture review process so that high-risk or strategically significant initiatives receive appropriate scrutiny without creating unnecessary bureaucracy for low-risk changes.
• Review material solution designs during discovery and qualification, before teams make irreversible technology or integration commitments.
• Issue clear review outcomes:
• Conformant.
• Conformant subject to stated conditions.
• Further design work required.
• Architecture exception required.
• Define and maintain the Architecture Decision Record standard, repository, ownership model, status lifecycle, and review process.
• Ensure Architecture Decision Records document the decision, context, options considered, rationale, consequences, risks, and responsible owner.
• Require accepted architecture decisions and reference patterns to be used during design, engineering review, and subsequent solution changes.
• Maintain a transparent architecture exception register showing:
• The nature and rationale of each exception.
• Associated risks and compensating controls.
• The accountable owner.
• The approved duration or review date.
• The remediation or exit plan.
• Recommend architecture exceptions to the Group CPTO for decision; the AI Solutions Architect does not approve exceptions to standards owned by the role.
• Track whether agreed architecture conditions and exception actions are closed within the committed timeframe.
• Escalate unresolved architecture or AI-by-design non-conformance to the Group CPTO through the defined governance process.

Cross-Lane Solution Design and Integration

• Work with Product Managers and Technical Product Managers during discovery and qualification to assess feasibility, architecture implications, technical dependencies, non-functional requirements, data implications, and AI-by-design conditions.
• Ensure Product and Technical Product Management teams receive timely and actionable architecture direction without transferring backlog ownership to the architecture function.
• Work with the CTO, IFI; CDS, IDD; Heads of Technology; and IDD functional heads to ensure solution designs conform to group architecture while leaving detailed engineering design and delivery accountability with the relevant delivery lane.
• Define common patterns for APIs, events, data contracts, identity, access, tenancy, integration, configuration, auditability, and observability.
• Prevent avoidable architectural fragmentation, unnecessary duplication, and product-specific forks where a configurable shared capability can meet the underlying need.
• Promote reuse of common services, models, integrations, and platform capabilities across IFI, IDD, and product lines.
• Assess the architectural impact of partner-bank, fintech, government, capital-provider, third-party, and ecosystem integrations.
• Ensure integration decisions account for failure handling, versioning, data ownership, service dependencies, operational support, and contractual service expectations.
• Facilitate resolution of cross-lane technical design conflicts and escalate only where the matter requires an executive trade-off or architecture exception.

Security, Data, Resilience, and Compliance Integration

• Partner with the CISO to ensure security-by-design and privacy-by-design requirements are incorporated into solution architecture from discovery.
• Ensure architecture reviews address identity and access, data protection, encryption, trust boundaries, threat exposure, third-party risk, logging, auditability, and secure failure modes.
• Work with the Chief Data Officer to ensure data architecture reflects approved ownership, classification, stewardship, lineage, retention, privacy, quality, and permitted-use requirements.
• Work with the Head of Data Science to ensure AI and model architecture supports the required validation, human oversight, fairness testing, monitoring, and lifecycle controls.
• Work with Infrastructure and DevOps to ensure solutions use approved infrastructure, deployment, secrets-management, environment, pipeline, and observability standards.
• Work with Customer Support Excellence to ensure architecture decisions support monitorability, supportability, operational runbooks, incident diagnosis, service-level commitments, and operational readiness.
• Define architecture requirements for availability, continuity, backup, disaster recovery, data recovery, graceful degradation, and failure isolation proportionate to the service's criticality.
• Ensure designs can accommodate applicable regulatory, data-residency, privacy, contractual, financial-services, and Sharia requirements across different markets.
• Support audit and regulatory review by ensuring material architecture decisions and AI controls are documented, traceable, and retrievable.

Architecture Modernization and Reuse

• Maintain an architecture-level view of legacy platforms, duplicated capabilities, integration complexity, and technology obsolescence.
• Develop transition recommendations that progressively align legacy solutions with Kifiya's target architecture without creating unnecessary delivery disruption.
• Establish criteria for modernization, re-platforming, refactoring, replacement, retirement, and controlled coexistence.
• Ensure modernization proposals clearly state business value, architectural benefit, delivery dependencies, risks, and total lifecycle implications.
• Identify opportunities to convert repeated implementation work into reusable services, reference implementations, templates, and platform capabilities.
• Measure the adoption and effectiveness of shared architecture patterns and use findings to improve the pattern catalogue.
• Ensure architecture-led modernization work is prioritized through the normal portfolio and backlog governance process rather than assigned directly to delivery teams.

Technology and Vendor Advisory

• Advise the Group CPTO and executive leadership on:
• Build, buy, or partner options.
• Platform and technology investment choices.
• AI and model-provider strategy.
• Vendor and integration architecture.
• Cloud, hybrid, and on-premise options.
• Multi-country expansion.
• Scalability and resilience implications.
• Technology concentration and lock-in risk.
• Architecture implications of major commercial commitments.
• Conduct or lead architecture due diligence for material technology vendors, acquisitions, partnerships, investments, and strategic customer engagements.
• Evaluate technology options against business outcomes, architecture fit, security, data use, interoperability, operational support, lifecycle cost, and exit feasibility.
• Present recommendations in language that allows executive and business stakeholders to understand the options, trade-offs, risks, and consequences.
• Provide recommendations and evidence; final build, buy, partner, investment, and architecture-exception decisions remain with the Group CPTO or the designated executive authority.

Architecture Capability and Practice Development

• Build architecture capability across IFI, IDD, Product, Infrastructure and DevOps, and other relevant teams through coaching, design clinics, reference implementations, and practical guidance.
• Establish an architecture community of practice that improves consistency while retaining engineering ownership within the delivery lanes.
• Coach Technical Product Managers, Heads of Technology, engineering leads, data leaders, and other design authorities on the use of reference architectures, Architecture Decision Records, non-functional requirements, and AI-by-design assessments.
• Make architecture standards easy to find, understand, apply, and test.
• Monitor relevant developments in AI, financial infrastructure, cloud, data, integration, security, and digital public infrastructure.
• Recommend controlled experiments where a new technology could materially improve Kifiya's products, delivery economics, risk posture, or ability to scale.
• Capture lessons from architecture reviews, incidents, delivery rework, and production performance and use them to improve standards and patterns.

Functional Authority and Role Boundaries

The AI Solutions Architect has the authority to:

• Define company-wide solution architecture and AI-by-design standards.
• Determine which approved reference architecture or pattern applies.
• Determine whether a material design conforms to approved architecture standards.
• Require architecture and AI-by-design conditions to be documented and addressed.
• Withhold architecture sign-off where material non-conformance remains unresolved.
• Chair the Architecture Review Board.
• Recommend an architecture exception to the Group CPTO.

The AI Solutions Architect does not:

• Set product vision, target outcomes, pricing, packaging, or commercial terms.
• Own customer or partner relationships.
• Own or order an IFI or IDD product backlog.
• Set sprint scope, release scope, or delivery sequence.
• Commit engineering delivery dates or allocate engineering capacity.
• Allocate DevOps capacity or own infrastructure and pipeline standards.
• Own detailed engineering implementation or code quality.
• Own model development, model validation, data fitness, or production model performance.
• Grant an exception to the architecture or AI-by-design standards owned by the role.
• Approve commercial, technical, and operational readiness on behalf of the accountable release authorities.
• Operate production platforms, lead incidents, or modify production code or infrastructure.

Key Outputs The role is expected to produce and maintain the following outputs:

• A current, approved, and communicated target and transition architecture for Kifiya's technology ecosystem.
• A version-controlled catalogue of reference architectures, integration patterns, AI solution patterns, and architecture guardrails.
• A documented AI-by-design standard and proportionate assessment method covering AI suitability, data use, controls, human oversight, evaluation, documentation, and monitoring.
• Completed AI-by-design assessments for applicable initiatives during discovery.
• Timely Architecture Review Board decisions supported by complete Architecture Decision Records.
• A maintained architecture decision repository and architecture exception register.
• Architecture conformance conditions incorporated into qualified delivery items.
• Architecture transition, modernization, and remediation recommendations routed into the appropriate portfolio and backlog process.
• Architecture due-diligence assessments for material vendors, platforms, partnerships, and investments.
• Architecture practice guidance, reusable templates, coaching materials, and reference implementations.

The effectiveness of the role will be assessed through measures including:

• Architecture conformance rate across material initiatives.
• Percentage of applicable initiatives completing an AI-by-design assessment during discovery.
• Adoption and reuse of approved shared components, services, and reference patterns.
• Architecture review and decision turnaround time against agreed service levels.
• Completeness and quality of Architecture Decision Records.
• Number, age, risk level, and closure rate of approved architecture exceptions.
• Reduction in integration rework and late-stage architecture redesign.
• Reduction in avoidable duplication and product-specific technical forks.
• Stakeholder confidence in the clarity, timeliness, and practicality of architecture guidance.
• Demonstrable improvement in the security, resilience, interoperability, supportability, and lifecycle efficiency of approved designs.

.

Qualifications and Experience Education

• Bachelor's degree in Computer Science, Software Engineering, Information Systems, Data Science, Artificial Intelligence, Computer or Electrical Engineering, or a closely related discipline.
• A relevant postgraduate qualification is advantageous but not mandatory where the candidate demonstrates equivalent depth of experience.

Professional Experience

• Substantial progressive experience, typically at least 10 years, in software engineering, data platforms, cloud, systems integration, solution architecture, enterprise architecture, or related technology disciplines.
• At least four years of experience in a senior solution, platform, data, cloud, or enterprise architecture role with responsibility across multiple products or technology domains.
• Demonstrated experience designing or governing production AI, machine-learning, data-driven decisioning, or advanced analytics solutions.
• Experience in financial services, fintech, banking technology, payments, credit, insurance, capital infrastructure, market infrastructure, or another regulated and transaction-intensive environment.
• Experience with distributed systems, API and event-driven integration, multi-tenant platforms, data architecture, identity and access, and high-availability services.
• Experience designing solutions across cloud, hybrid, and on-premise environments; AWS experience is preferred.
• Experience supporting multi-country, partner-led, or configurable platform deployments is strongly preferred.
• Demonstrated experience establishing architecture governance, reference architectures, design-review processes, Architecture Decision Records, and controlled exception management.
• Experience influencing senior technical and business stakeholders without relying on direct line authority.
• Experience conducting build-versus-buy assessments, vendor architecture reviews, technical due diligence, or strategic technology evaluations.
• Experience in Africa or comparable emerging-market environments is advantageous.

Technical and Governance Knowledge

Strong understanding of:

• AI and machine-learning solution architecture.
• Data and model lifecycle dependencies.
• Responsible AI and AI risk management.
• Human-in-the-loop and human-oversight design.
• Model integration, inference, monitoring, and fallback.
• Generative and agentic AI architecture.
• Data governance, privacy, lineage, and permitted use.
• Distributed systems and integration architecture.
• API, event, and data-contract design.
• Cloud-native, hybrid, and on-premise architecture.
• Security, resilience, observability, and operational readiness.
• Multi-tenancy, configuration, and regional deployment.
• Technology lifecycle, modernization, and technical debt.
• Working knowledge of recognized AI governance, risk-management, security, and architecture frameworks and the ability to adapt them pragmatically to Kifiya's context.
• Ability to move between executive-level architecture decisions and sufficient technical detail to test whether a proposed solution is credible and implementable.
• Relevant cloud, architecture, security, data, or AI-governance certifications are advantageous but are not a substitute for demonstrated architecture judgment and delivery experience.

Core Competencies

• Enterprise and ecosystem thinking: Understands how products, data, AI, platforms, infrastructure, operations, security, and commercial outcomes interact as one system.
• AI-by-design judgment: Distinguishes genuine AI opportunities from inappropriate use and applies proportionate controls to the full AI lifecycle.
• Architectural depth: Understands complex solution trade-offs across applications, data, integration, AI, infrastructure, security, and operations.
• Pragmatic governance: Creates sufficient discipline, evidence, and consistency without allowing governance to become unnecessary bureaucracy.
• Functional authority: Influences and sets standards across teams without depending on direct reporting authority.
• Decision quality: Makes timely, evidence-based, and explainable decisions under uncertainty.
• Business and financial-infrastructure acumen: Connects architecture decisions to customer value, revenue, risk, regulatory obligations, scalability, and lifecycle economics.
• Risk awareness: Identifies security, privacy, data, AI, integration, resilience, vendor, and operational risks early enough to influence design.
• Structured problem solving: Breaks complex cross-system issues into clear choices, dependencies, trade-offs, and consequences.
• Executive communication: Presents technical issues in concise business language while retaining the integrity of the underlying analysis.
• Collaboration across seams: Works effectively with Product, Technical Product Management, engineering, data, security, infrastructure, operations, risk, compliance, and business stakeholders.
• Courage and integrity: Is prepared to withhold sign-off, document disagreement, and escalate material risk while remaining constructive and solution-oriented.
• Learning agility: Keeps standards and patterns current as AI, financial infrastructure, regulation, and technology evolve.
• Operational discipline: Maintains reliable decision records, standards, registers, actions, and follow-through.