Josué Avuh
Solution Architect
Cloud · AI · Data
“I think in systems, speak in business, and ship in code.”
I architect intelligent systems at the intersection of Cloud, AI, and Data — for organizations where precision is non-negotiable.
About
Solution Architect based in Benin, operating internationally with teams across France, Mauritius, and Morocco. I design and ship Cloud, AI, and Data architectures for high-stakes organizations — from POC through validated architecture to production systems.
My technical work drives two levers. First, I've trained consultants now deployed at BPCE, Hermès Paris, SwissLife, Crédit Agricole, MGEN, and Groupama — and I run the technical oversight of those engagements in parallel to secure delivery quality. Second, I've led multiple architecture-validation POCs that served as the technical green-light before structuring missions officially kicked off.
Master's in Information Systems (AUB, 2024). Multi-cloud — AWS, Azure, GCP, Alibaba. AI-native — I integrate AI as a measurable operational lever, not a buzzword.
Expertise
Cloud Architecture
From POC to production, at scale.
AI & Agent Engineering
AI as a co-pilot, not a gimmick.
Data & Business Intelligence
From raw data to strategic decisions.
Architecture
Architectures built from real projects — from first sketch to deployable system.
NeoCredit — Real-Time Credit Intelligence Platform
Credit application data sat inside transactional systems with no dependable CDC backbone, limiting real-time visibility for BI, reporting, and AI use cases.
I designed a CDC-driven multi-cloud data platform where AWS RDS changes are captured by Debezium on EC2, streamed through Kafka, processed by AWS Glue Spark, then delivered to both BigQuery/Vertex AI/Looker Studio and Snowflake/dbt/Power BI.
Real-time credit intelligence · Dual analytical backbone · Production-ready foundation for BI, reporting, and AI
Enterprise AI Copilot — Azure AI Foundry
Knowledge workers wasted hours searching scattered docs, while safety, audit, and traceability gaps blocked LLMs from moving past POC into production.
I designed a multi-agent platform on Azure AI Foundry — Foundry Agent Service orchestrates tool calls across Azure OpenAI (with Claude via Foundry), AI Search RAG over governed Blob storage, Content Safety guardrails, Cosmos DB session memory, and full Purview audit lineage.
Search time ÷2 · 100% compliance-grade audit trail · 70% of repetitive tasks delegated safely
Event-Driven Digital Backbone
A monolith with fragile point-to-point integrations — cascading failures, three months to onboard a new partner, no observability on in-flight events.
Domain services publishing to an MSK Kafka backbone with schema registry discipline, Flink stateful processing, DLQ replay for error recovery, and an AI ops copilot triaging anomalies on curated Snowflake/dbt signals.
MTTR ÷4 · Partner onboarding from 3 months to 2 weeks · 99.95% event delivery
Certifications
AWS Solutions Architect Associate (SAA-C03)
Microsoft Azure AI Engineer Associate (AI-102)
Microsoft Business AI Solution Architect (MB-280)
Let'stalk
Technical challenge, consulting mission, or strategic collaboration — let's talk.