Route log
An unusual route to engineering
2011 – 2024
Owner & Manager
Nick Kalas Winery
Ran a production business end to end — operations, planning, equipment, seasonal crews, customers. Thirteen years of owning outcomes.
2024 – 2025
Truck Dispatcher
Iron Wing Dispatching
Dispatched US freight — brokers, drivers, rate confirmations, settlements. Lived the operational pain that became Kelévo's spec.
2024 – Present
Founder & Founding Engineer
Kelevo
Self-taught via The Odin Project, then built a commercial TMS SaaS solo: product, architecture, code, infra, launch.
Featured system
demo launching soonKelévoTransportation Management System — SaaS
Dispatching, fleet & asset management, OCR document intake, and settlements — the ops backbone for small and mid-size trucking companies. Built solo, zero to launch.
- #10428Atlanta, GA → Dallas, TXDELIVERED
- #10431Memphis, TN → Chicago, ILIN-TRANSIT
- #10436Denver, CO → Phoenix, AZTENDERED
- #10442Kansas City, MO → Omaha, NEPLAN
Dispatch & loads board
Full load lifecycle from tender to delivery, driven by an explicit state machine — no invalid transitions, ever.
Fleet & asset management
Trucks, trailers and drivers with their own lifecycles, assignments and per-org numbering.
AI/OCR document intake
Rate confirmations imported from PDFs — OCR, validation, and structured extraction into loads.
Settlements & statements
Owner-operator settlements with branded, immutable PDF statements emailed on issue.
Multi-tenant core
Org-scoped data isolation enforced at the service layer, with plan- and role-based capability gating.
Stripe billing
Trial-to-paid subscription lifecycle wired into plan entitlements.
Case study
Building Kelévo
The problem
Small trucking companies run million-dollar operations on spreadsheets, phone calls and inbox archaeology. As a dispatcher I lived it daily: every load means a rate confirmation to parse, a driver to assign, status calls to brokers, and a settlement to reconcile — all manual, all error-prone. Existing TMS products are priced and designed for large fleets. Kelévo is the system I wished I had at the dispatch desk.
Architecture
- ClientsWeb app (React)Pro app (React)Mobile (planned)
- APINode.js / ExpressZod-validated routesPlan+role gating
- Domain coreShared TS domain packageLifecycle state machinesOrg-scoped services
- Data & integrationsPostgreSQL + PrismaCloudflare R2 documentsStripe billingAI/OCR intake
- InfraDockerGitHub Actions CI/CDnginx edgeVPS + EC2
Hard decisions
A shared domain vocabulary
One TypeScript domain package owns enums, types and DTOs across backend and frontends — with CI drift-guards so the database schema, API and UI can never silently disagree.
Explicit lifecycle state machines
Loads, trucks and trailers move through authoritative transition tables, not scattered if-statements. Invalid transitions are unrepresentable, and the docs mirror the code.
Multi-tenancy at the service layer
Every query is org-scoped at the service boundary — tenant isolation is a structural guarantee, not a convention reviewers must catch.
Document-heavy by design
Freight runs on PDFs. Rate confirmations flow in via email + OCR into structured loads; settlements flow out as branded, immutable PDF statements.
Outcomes
- A commercial, multi-tenant TMS SaaS taken from zero to launch by one person.
- A strictly-typed monorepo with a single consolidated CI gate: format, lint, typecheck, tests, typing-drift guards.
- Real operational workflows — dispatch, fleet, documents, settlements, billing — not a demo app.
Stack & skills
What I work with
Languages
Product & Frontend
Backend
Cloud & DevOps
Execution
Contact
Let's talk
Open to founding-engineer and senior full-stack roles — especially where logistics domain depth matters.