LK Work with me

LK Lkhadorj Lkhagvasuren

I learned this business the hard way: cold calls, loads that fell through, paperwork at 2 a.m. Today I run a carrier and a licensed brokerage out of Houston on systems I built myself. Now I build them for other companies.

AI implementation, applied research and new companies, from the first filing to the first customer.

I learned the business on the lot.

Before I hauled cars, I sold them: CarMax, then acquisitions at Russell & Smith Ford. The lesson that stuck: buyers don't pay for promises. They pay for certainty.

Best single deal: $22,000 in gross.

Then I put the trucks on the road.

LK Auto Transport runs car haulers from Houston to Wisconsin and back, and everything in between. 300+ cars moved, every mile and dollar tracked in software I built.

  1. Houston
  2. Shreveport
  3. Arkansas
  4. Missouri
  5. Illinois
  6. Wisconsin

USDOT 4528474, MC 1795270.

Then I started the network.

Chinggis Logistics is a licensed, bonded auto-transport brokerage. We connect dealers, auction buyers and families with carriers we trust, and every carrier clears our own multi-step vetting before it touches a car.

  1. Company formed
  2. Applied for federal broker authority
  3. Licensed by FMCSA
  4. Taking bookings online

MC-73664766, backed by a $75,000 broker surety bond.

At night, I build the software.

No dev team and no budget for one. I took what years in business, sales, tech and logistics taught me, built the systems myself, and pushed each one a step further with AI. Four of them run the companies today.

LK Tracker's Loads screen on a phone: trips out of Houston with each load's route, cars, status and profit
Real screen, sample data

LK Tracker

The carrier's whole back office in one app, on a phone in the cab or a desk at home.

Purpose
Two spreadsheets were running a trucking company, and losing. I replaced them with one app for loads, trips, expenses, broker payments, fuel taxes, compliance dates and a monthly P&L.
Stack
TypeScript, Fastify, SQLite, React as an installable phone app, Tailwind, an AI API that reads rate confirmations, connections to banking apps, cloud computing.
Results
  • 399 commits since May 2026; it runs the carrier in production.
  • 143 API routes, 43 tables, 118 test files.
  • Drop in a stack of rate confirmations and AI reads every one. Nothing saves until you approve the preview.
  • A full accounting audit against production on July 29, 2026 found every payment split exact.
The Chinggis broker app's Today page: what needs attention, money owed both ways, and where every lead and load stands The Loads board: loads moving from Booked through Vetting, Dispatched and Delivered to Closed, each with its margin Invoices: what each customer was billed, what came in and what is still open
Real screen, sample data

Chinggis broker app

From the first call to the paid invoice, in one place.

Purpose
Runs the brokerage's whole day: leads, quotes, booking, dispatch, rate confirmations, invoices, carrier payables and reports, with roles for owner, agents and viewers.
Stack
TypeScript, Fastify, SQLite, React, secure card payments, Google APIs, server-made PDFs, built-in communication systems.
Results
  • First commit July 16, 2026; leads, loads, invoices and reports live by August 26, with 119 backend tests.
  • A carrier that fails vetting cannot be dispatched. The app stops it, not a reminder.
  • Instant public quotes across 55 makes and about 700 models.
Ordo's 3D map: the LK Auto Transport and Chinggis Logistics districts, each department a building, linked to LK's tower
Real screen, sample data

Ordo

Agents plan and do the work. I make the decisions.

Purpose
Named for the khan's headquarters, where every report arrived and every order went out. One cockpit for both companies: scheduled agents brief me, chase invoices, watch deadlines, research and check each other's work, and Ordo shows only what needs my yes.
Stack
Python, cloud computing, three.js for a walkable 3D office, Git and YAML as its memory, voice, scheduled agents that reach me on my phone.
Results
  • Live October 7, 2026: 10 scheduled routines and 4 always-on agents.
  • 139 routine runs in its first two weeks.
  • At most three signals a day and one morning briefing. No outreach goes to a customer without my tap.
Wednesday, October 21 5:42
Scoutnow
📋 Text the broker (tap to copy) Hi, this is LK Auto Transport. We can take your 3 cars Houston to Chicago, pickup 10/22, $650/car. Can you send the dispatch?
Scoutnow
🚛 New load: Houston, TX → Chicago, IL $1,950 ($650/car), 3 cars. Pickup on/after Oct 22.
Scout2m ago
Watching your saved search. Checked 2 minutes ago.
Sample data

Scout load alerts

New loads, the moment they post.

Purpose
Watches a saved load-board search and puts every new load on my phone the moment it posts, in place of a paid alert service.
Stack
A browser extension in TypeScript, Fastify, SQLite and cloud computing, with alerts that land on your phone.
Results
  • Checks about every 2 minutes, one alert per new load, with quiet hours and a rate floor.
  • Each alert carries a ready-to-send text for the broker.
  • Tells me when it gets signed out or stops seeing loads, once per bad spell.

Fraud is the problem I can't put down.

My experience as a JPMorgan Chase banker showed me how fraud really gets in: rushed people, skipped steps, and identities nobody double-checks.

Auto transport runs on the same blind trust. A broker hands a car worth tens of thousands of dollars to a carrier found minutes earlier on a load board. Fraudsters hijack real carriers' registrations, spin up clean-looking shells and re-broker loads they never haul. It's what keeps me up at night.

116,455 of the 213,351 active FMCSA-registered vehicle haulers report a single truck.

FMCSA Company Census File, October 2026. Cargo class is self-reported.

Gerege Labs

My research on carrier fraud: link carrier identities across competing brokers without any of them sharing raw data, track the federal registry day by day, and give every alert a reason a dispatcher can read.

NSF SBIR Project Pitch submitted, October 2026.

A gerege was the gold tablet a rider carried for safe passage across the Mongol Empire: trust you could hold in your hand.

LK in a navy suit, smiling

I'm LK.

I studied Supply Chain and Logistics Technology at the University of Houston, where my research team made the top 20 at the BreakBulk Americas competition. Since then I've sold cars, worked fraud procedures at a bank, put trucks on the road and built a brokerage from nothing. Every mistake along the way became a system. I speak English and Mongolian, and I build every day.

Bring me in for

AI implementation
I've put agents to work inside two real companies, with the guardrails and approval steps that keep the owner in charge. I'll do the same in yours.
Applied research
Fraud, trust and logistics data, written up to the standard grants and partners expect.
Company building
From the first filing and federal authority to the software and the first customers. I've done it twice this year.

Tell me what you're building.

Email me Book a 20-minute call LinkedIn
What do you need?