Avery LindqvistExample: Avery and Larkspur are made up. Make your own

Prepared for the Senior Backend Developer role at Larkspur Logistics

I've already built the dispatch board your customers live in

At a regional trucking company I replaced three spreadsheets and a whiteboard with a live dispatch board that checks hours-of-service rules before every assignment. Larkspur's Dispatch Platform team builds that same kind of tool for whole fleets, and I'd like to help.

Five production tools shipped for operations teams over four years.

Avery Lindqvist for Larkspur Logistics1 / 8

What the role needs, and where I've done it

Production backend services in TypeScript on Postgres
Dispatch Board runs on Next.js, TypeScript and Postgres for four depotsDispatch Board
Real-time systems over WebSockets
Live updates keep every dispatcher on the same boardDispatch Board
Validating assignments against hours-of-service rules
The board checks remaining driver hours before it saves an assignmentDispatch Board
Working with non-technical users and measuring impact
Built with dispatchers and billing staff, with results measured after launchDispatch BoardInvoice ExtractorClinic Scheduler
LLMs for structured extraction from documents
Invoice Extractor reads carrier PDFs into a fixed schema, tested on 400 invoicesInvoice Extractor
Avery Lindqvist for Larkspur Logistics2 / 8

Northwind FreightInternal tool2023 – 2024

Dispatch Board

Live load assignment for dispatchers

The closest thing to your product I've shipped: load assignment with hours-of-service checks, used daily.

What it does

  • Drag-and-drop load assignment with hours-of-service checks
  • Late-pickup warnings 2 hours ahead
  • Live updates across all four depots
Results
  • Late pickups down 31% in the first quarter
  • Used by all 22 dispatchers daily
Used by
Dispatchers and Depot managers
Built with
Next.js, TypeScript, Postgres, WebSockets
Avery Lindqvist for Larkspur Logistics3 / 8

Northwind FreightAI pipeline2024

Invoice Extractor

LLM reading of carrier invoices

Your bonus requirement, in production: LLM extraction with an evaluation set and a human review queue.

What it does

  • Schema-constrained extraction from PDFs
  • Automatic rate comparison
  • Review queue for mismatches
Results
  • Recovered $84,000 in overbilling in year one
  • Extraction accuracy 96% on a 400-invoice test set
Used by
Billing team and Finance
Built with
Python, FastAPI, Claude API, Postgres
Avery Lindqvist for Larkspur Logistics4 / 8

A clinic networkProduction app2022

Clinic Scheduler

Waitlist that fills cancellations

Another operations team, another legacy system, measured the same way.

What it does

  • Priority-ordered SMS offers
  • Books the first confirmed reply
  • Audit log for front-desk staff
Result
Unused cancelled slots down from 40% to 12%
Used by
Front-desk staff and Clinic managers
Built with
Node.js, TypeScript, Twilio, MySQL
Avery Lindqvist for Larkspur Logistics5 / 8

More projects

  • tidy-csv

    A command-line tool that fixes the messy CSV exports operations teams live with

    Open
Avery Lindqvist for Larkspur Logistics6 / 8

How I work

  1. 1Sit with the team
  2. 2Find the slowest step
  3. 3Ship a small version
  4. 4Measure
  5. 5Iterate

Frontend

  • React, Next.js, TypeScript
  • Tailwind, accessible UI
  • Playwright tests

Backend & data

  • Node, Python, FastAPI
  • Postgres, Redis
  • Queues and scheduled jobs

AI & delivery

  • LLM extraction and classification
  • Evaluation sets
  • Docker, CI/CD
Avery Lindqvist for Larkspur Logistics7 / 8

If the Dispatch Platform team is still hiring, I'd welcome 20 minutes to walk you through the board and what I'd build next.

Avery Lindqvist

Full-stack developer, Remote