Northbeam Capital — Ledger Modernization
Rebuilt a 14-year-old core ledger into an event-sourced service mesh, cutting nightly reconciliation from 9 hours to 11 minutes.
98%
Faster batch
$3.2M
Annual savings
Selected Work · 2019 – 2025
Engineering measurable outcomes for ambitious teams.
A curated index of case studies where sharp strategy, modern engineering, and tightly-scoped design translated into revenue, retention, and runtime performance our clients can feel on a Monday morning.
84+
Shipped Engagements
$2.1B+
Client ARR Supported
6yr
Avg. Partnership
96%
Retention Rate
Case Studies
Selected work, shipped with intent.
Each engagement below was scoped to a business outcome — not a feature list. We work with operators who measure twice and ship once, and we treat their metrics as our north star.
Cobalt Analytics — Onboarding Revamp
Redesigned a 17-step setup into a 4-step guided activation, driving 3.4x lift in week-one retention for B2B accounts.
3.4x
W1 retention
+47%
Activation
Verum Health — Clinical Triage Engine
Built an LLM-assisted triage copilot for nurse practitioners, with strict PHI boundaries and a 220ms median response budget.
62%
Time saved
0
PHI incidents
Meridian Freight — Multi-Region Cutover
Migrated a 9-region monolith to a cell-based architecture on AWS, with zero-downtime cutover and full observability parity.
71%
Lower latency
99.99%
Uptime SLA
Quorum Studios — Creator Companion App
Shipped a React Native creator tooling app in 11 weeks with offline-first sync, used daily by 38k creators across 14 markets.
4.8★
App Store
38k
DAU
Halcyon Retail — Demand Forecast Pipeline
Replaced a brittle Excel demand model with a streaming feature store + ML pipeline, lifting forecast accuracy from 64% to 89%.
89%
Accuracy
$11.4M
Waste avoided
Before & After
A representative shift, in numbers.
Verum Health's clinical triage copilot, twelve weeks in. Baseline measured against the prior rule-based workflow over a 30-day parallel run.
Before — Rule-based triage
Avg. time to triage
14m 22s
Clinician satisfaction (CSAT)
3.1