About
I’m Ryan de Melo, a technology executive and hands-on builder. For about eighteen years I’ve built and scaled AI/ML platforms, data infrastructure, and cloud-native systems in production, across e-commerce, payments, fintech, supply chain, and media.
Right now I’m founder and CTO of an enterprise AI venture, building GenAI infrastructure for financial services and industrial operations. That means I’m back in the code: RAG pipelines that resolve real accounts-receivable discrepancies, agentic workflows over financial systems with humans in the loop, and multimodal inspection that goes from voice and image straight to action in the field.
What I’ve built
A few things I’m proud of, and the kind of work I write about here:
- A B2B OpenAPI platform taken from $0 to $1.1B in annual revenue. A new business line, not an extension of an old one.
- ML systems serving 550M+ recommendations a day, lifting conversion 18% and order value 12%.
- The payments, BNPL, credit scoring, and fraud platforms behind $22B+ in commerce at 80M+ monthly users, supporting 4.4x growth.
- A self-service ML platform (feature store, model serving, vector database) used by 20+ product teams, cutting deployment cost 26%.
- A multi-cloud migration of 380+ services and 3PB of data across four providers in 16 months, ahead of schedule.
- A post-merger technology integration that harmonised ten overlapping systems in nine months with zero customer disruption.
I’ve also done the parts that don’t make highlight reels: standing up compliance for an IPO path (ISO 27001/27701, SOX 404, PCI-DSS), meeting data residency and sovereignty rules under regulators like Indonesia’s OJK, and negotiating a $350M hyperscaler partnership.
How I’ve worked
I’ve led distributed engineering, applied-AI, and product organizations of 500+ across Singapore, Indonesia, India, China, and the US. I grew one org from 120 to 500+ people and built the bench behind it, moving ICs to managers and managers to directors. I’ve presented technology strategy and risk to boards, CFOs, and pre-IPO investors, and I’ve sat on the technical side of M&A diligence.
The throughline is the same at every size: build platforms that stay reliable as traffic, data, and headcount grow, and build the teams that run them well.
What I write about
This blog is where I think out loud about the work:
- Architecture — real decisions under real constraints, and why one option beat the others
- Scale — what actually changes when traffic, data, or headcount get large
- Applied AI — RAG, agents, and MLOps in production, not in slideware
- Leadership — restructuring teams, raising the bar, and the unglamorous middle of delivery
No fluff, kept short and direct. If a post saves you a week of pain, it did its job.