AI Engineer & Solution Architect building LLM-powered production systems
I'm Richard Emijere, an AI Solution Architect and Senior Software Engineer with 5+ years delivering production-grade systems and 2+ years leading GenAI implementations in fintech. I specialise in multi-agent AI systems, model routing, RAG pipelines, and MLOps on AWS and Azure. Currently building AI banking platforms at FCMB Group that process 100K+ daily transactions.
Work Experience
FCMB Group
Senior Software Engineer - AI Systems
Architected and own a production-grade AI platform integrated with core banking APIs, processing 100K+ daily financial transactions at 99.9% uptime. Engineered cost-aware model routing across OpenAI, Anthropic, and Gemini reducing inference costs by 30%. Led delivery of an agentic AI banking assistant using RAG and LLM orchestration, automating 60% of customer support interactions. Built async WebSocket middleware cutting end-to-end latency by 60%.
Qore
Software Engineer - System Integration & Backend
Developed backend engineering for Pryme middleware powering automated Card Vending Machines (CVMs) and virtual card issuance across 6 major Nigerian banks. Engineered resilient core banking integrations to automate real-time account validation, transaction posting, and debit bypassing - eliminating manual reconciliation and reducing transaction drop-offs. Designed event-driven architecture with RabbitMQ for asynchronous branch card allocation and decentralised inventory tracking. Maintained and scaled C#/.NET and TypeScript microservices on Azure Kubernetes Service (AKS) with structured logging and robust CI/CD pipelines.
Microsoft
Software Engineer (Associate)
Led architecture and delivery of BranchOps - an LLM-powered CI/CD intelligence platform - as tech lead at Microsoft Global Hackathon, improving deployment efficiency by 35%. Built high-performance RESTful APIs (Python/FastAPI, C#) handling millions of daily requests at sub-100ms response times. Engineered ETL pipelines processing 500K+ daily records with Azure Synapse Analytics, reducing latency by 30%.
Integral Information Technologies
Fullstack Developer
Built a government Electronic Physical Planning System (EPPS) with React, Next.js, and TailwindCSS, digitising end-to-end approval workflows for 10,000+ concurrent users at 99.8% uptime. Developed real-time analytics dashboard integrating backend RESTful APIs on AWS, improving Core Web Vitals (LCP, TTI) by 25% through code-splitting, caching, and indexed PostgreSQL schema design.