Most software problems aren't hard because of the code.

They're hard because of everything around it: data you're not allowed to move, rules that differ for every customer, numbers that have to be defensible to someone who isn't in the room. That's the work I take.

Decisionsten years of making themSources that disagreesalesforce cpq vs. general ledgerData you can't migrateyears of live clinical recordsRegulation you can't negotiateclinical identity, per jurisdictionMoney that has to tie outpro-rata creator payoutsRules differ by customercustomer-defined work categoriesSurvives diligenceboard & sponsor reportingClinician trustbehaviour they already knowPayouts that clearreconciled against revenueExtensible without a rewritethe model outlasts me
Constraints that disagreeOutcomes that have to hold

Ten years across regulated healthcare, enterprise multi-tenancy, LLM product work, and financial reporting under private-equity scrutiny.

Epic Systems · WorldQuant
Founding engineer, CuroWork

01 Selected work

Four systems, built end to end.

2023 - 2026

Multi-tenant workforce visibility

CuroWork

CuroWork helps finance and back-office teams plan work, see bottlenecks, and compare expected effort with what actually happened. I owned the platform end to end, from its multi-tenant data model and bidirectional calendar sync to production LLM features. In one customer workflow, month-end close fell from 20 days to 12 in three months.

  • Next.js
  • TypeScript
  • PostgreSQL
  • Supabase
  • LLM features
14 concurrent B2B contracts
Read the case study: CuroWork
2016 - 2019

Clinical scheduling & patient history

Epic Systems

I led the Patient History team through a full rewrite, then took on Provider/Nurse Scheduling: the primary clinical entry point for Epic's Ambulatory users. There is no maintenance window where a hospital stops seeing patients.

  • Healthcare
  • Data modeling
  • Multi-tenancy
  • Team leadership
Solved critical upgrade issue for 30%+ of the customer base via a simple override mechanism
Read the case study: Epic Systems
2026 - present

A PE-acquired B2B SaaS company

Reporting infrastructure

After an acquisition, the reporting a company runs on internally is rarely the reporting its sponsor needs. The business logic lived in Salesforce CPQ, not in anything resembling a data model.

  • Snowflake
  • Tableau
  • Salesforce CPQ
  • SQL
One governed source of truth for ARR, P&L, and balance sheet
Read the case study: Reporting infrastructure
2020 - 2022

iOS podcasting & creator payouts

Podder

Paying creators pro-rata by listening duration is a metering and settlement problem, not a feature. Getting the money right is where consumer products with revenue sharing usually break.

  • Swift / iOS
  • Node.js
  • MongoDB
  • Angular
  • Payments
Native iOS product shipped to the App Store
Read the case study: Podder
Core stack

TypeScript · React · Next.js · Node.js · Angular · Swift / iOS · C# / .NET
PostgreSQL · Supabase · MongoDB · Snowflake · Tableau · AWS · Vercel
LLM integration · Structured extraction · Embeddings & semantic retrieval

But the systems above didn't get hard because of the language they were written in. They got hard because several of these were true at once.

CuroWork
Epic
Reporting
Podder
Regulated / clinical data
no
yes
no
no
Rules differ per tenant
yes
yes
no
no
Live system, no downtime
yes
yes
yes
no
Money has to reconcile
no
no
yes
yes
Third-party APIs disagree
yes
no
yes
yes
Owned end to end, alone
yes
no
yes
yes

02 About

Ten years, mostly on hard constraints.

Kai Russo

I'm an engineer in Chicago. I spent three years at Epic Systems leading teams on Patient History and Provider/Nurse Scheduling, left to build products as a founding engineer, and now work independently.

I work best where the constraints are real: compliance, scale, a data model that can't be rolled back, or a number that has to survive a diligence process. Earlier, I built equity-allocation models for WorldQuant across US, European, and Asian markets, and was selected as one of the top three US consultants for the 2015 World Summit.

B.A. Mathematics & Computer Science, minor in Economics @ Boston University.

CareVeritas

Risk scoring and simulation helping financial advisors model client exposure to family-caregiving costs.

WorldQuant

Equity-allocation models on fundamental and price/volume data, deployed out-of-sample across three continents.

03 Contact

Tell me what's stuck.

If you're deciding whether something is worth building, that's a good conversation to have early, before the estimate, not after.

mail@kairusso.com Chicago, IL - remote