I design backend platforms that stay fast and reliable as they scale — payment rails, conversational-banking AI and event-driven risk engines serving 12M+ users.
I’m a principal-level engineer who has spent 9+ years where reliability meets craft — payment rails, conversational-banking AI and cloud-native platforms handling millions of users daily across fintech, banking, energy and SaaS.
Today I work on designing systems, defining technical standards, shaping roadmaps and mentoring engineers. I care most about building things that outlast any single sprint.
A selection — the work I keep coming back to.
A social travel platform connecting travelers to share trips, meet safely and build real experiences. Designed and built end to end — product strategy, architecture and full-stack development.
Re-architected a risk-processing platform for North America’s top energy firms — from failing MSSQL stored procedures to event-driven Azure Container Apps. Horizontal scale across 100–250 containers, processing 10M+ assets per run with built-in fault recovery.
A customer-facing banking AI handling account opening, transfers, loans and support through natural conversation. Built on Microsoft Bot Framework and Azure Cognitive Services, serving up to 500K users for one of Africa’s largest banks.
High-concurrency, low-latency USSD APIs powering mobile-wallet services for 12M customers — balance enquiries, transfers and financial inclusion at scale. Redis caching for throughput under peak load, lifting transaction success 10%.
A multi-tenant payment-as-a-service platform serving 100+ clients with multi-currency accounts for payouts, collections and exchange. Integrated global providers: Clearbank, Currencycloud, Stripe, Mastercard, Plaid and Providus.
A single MSSQL stored-procedure engine ran risk calculations for some of North America’s largest energy firms. As portfolios grew it buckled — jobs timed out, a single failure restarted the whole run, and work couldn’t scale across machines.
Event-driven fan-out instead of scaling the database vertically — throughput grows with containers, not hardware.
Container Apps over serverless Functions for long-running compute and predictable cost at peak scale.
Accepted eventual consistency for parallelism and fault isolation — each asset processes and retries on its own.
Nine years across fintech, banking and energy — tap any role to expand.
Principal work is multiplier work — the standards, reviews and systems that let a whole team move faster and break less.
Standards, docs and reviews outlive any single sprint.
Measure first, then refactor where it moves the needle.
Monitoring, fault recovery and 99.9% uptime designed in, not bolted on.
The best architecture serves the user and the business behind it.