What I’ve built. What it delivered.
Projects across healthcare and fintech — each one a high-stakes architectural decision with a measurable result. Details anonymized.
From $350K a year to $50K, processing billions of genomic sequences in the cloud.
A healthcare company was spending $350,000 a year on dedicated specialized servers to process their genomic data — and the infrastructure couldn't scale without proportionally scaling the cost. The founders brought me in to evaluate whether cloud was a viable alternative. I redesigned their entire data processing architecture on AWS, replacing the fixed server infrastructure with a cloud-native solution purpose-built for genomic workloads at scale. The system went to production processing billions of sequences a month at $50,000 a year — an 85% cost reduction — while handling significantly more volume than the dedicated servers ever could.
"The dedicated servers couldn't scale without scaling the cost. AWS HealthOmics changed that equation entirely."
Freed up $300,000 in annual infrastructure spend. That capital went back into product, hiring, and growth instead of servers.
1.6 million financial forms a year. One intelligent pipeline. $100K.
A fintech company was processing 1.6 million financial documents a year through an aging on-premise system with no visibility into what it was actually costing per document. The director brought me in to design a cloud-native alternative before committing to another on-prem cycle. I built a fully automated document processing pipeline on AWS that extracted and structured data from each form at scale, from raw document intake to clean, usable output. The system went to production processing 1.6 million documents a year at $100,000, with full cost transparency the previous system never provided.
"For the first time, we had a clear cost-per-document figure. On-prem, that number was invisible."
Built a pipeline where cost per document decreases as volume grows. The economics get better the more they scale.
Surgical data in 10 minutes instead of hours, and a 5-year on-prem contract avoided.
A healthcare technology company was days away from signing a 5-year on-premise server contract when the director brought me in for a second opinion. Their systems were generating operational data that doctors needed to review — but it was taking hours to arrive due to a poorly designed processing layer. I redesigned the data pipeline on Snowflake, removing the bottleneck entirely. Data that took hours to reach the clinical team now arrives in 10 minutes. The 5-year contract was never signed.
"Hours to 10 minutes. For surgical data, that's not just a technical improvement."
Turned a multi-hour data delay into a 10-minute delivery window, and avoided locking $millions into a 5-year infrastructure commitment that would have constrained the business for years.
Ready to know what’s actually in your next deal?
Every engagement above started with a 30-minute conversation. No commitment, no pitch — just a straight discussion about the technology and whether an assessment makes sense for your timeline.
Book a free 30-min call →