I'm often asked why a former Google Software Engineer with experience building machine learning systems and large-scale infrastructure would leave Silicon Valley to build software for the Texas hemp industry. The honest answer? Because the most interesting problems aren't always the most fashionable ones.
The Problem That Pulled Me In
Texas is home to thousands of hemp retailers — 7,289 businesses across more than 13,000 licensed retail locations. It's a real industry, run by real business owners, operating under regulatory requirements that keep getting more demanding.
Yet most of those businesses run on spreadsheets, paper binders, and generic tools that were never designed for a regulated product. Certificates of Analysis live in email attachments. Age verification is a judgment call. When an inspector shows up, getting audit-ready means a scramble.
Texas rewrote its hemp rulebook with 25 TAC Chapter 300, the DSHS regulations that took effect in March 2026 — licensing fees, product rules, and active inspections. This isn't a "move fast and break things" market. It's a market that needs dependable infrastructure.
That's when I knew someone needed to build the hard infrastructure first — someone who could apply serious engineering discipline to an industry that desperately lacked it. As a software engineer with a Mathematics degree from UNT and a background in machine learning systems, I had the exact skillset this problem demanded.
The Contrarian Decision: Build Before Launch
Most startup advisors would tell you to launch an MVP in two weeks, get users, iterate fast. I did the opposite.
I spent 14 months building before launch — at zero salary. Fourteen months writing code, architecting systems, and building compliance-first infrastructure designed to hold up in one of the most demanding regulatory environments in the country.
People thought I was crazy. Why spend 14 months building when you could be acquiring customers?
Here's why: in regulated industries, you don't get second chances. One compliance failure can shut a business down. One age verification miss can carry serious penalties. One expired license can destroy credibility overnight.
I wasn't building another app. I was building infrastructure for businesses whose survival depends on getting compliance right. The stakes were too high to get wrong.
What We Built in 14 Months
I built the initial platform myself, then led a 7-person AI-augmented team through the pre-launch build. Together we shipped what most startups would need a much larger engineering team to create:
A Suite of Production Applications
- Vendor Portal: Order fulfillment, inventory management, compliance workflows, analytics
- Customer App: Product browsing, ordering, real-time tracking
- Courier Portal: Route optimization, proof of delivery, earnings tracking
- Operations Dashboard: Real-time monitoring, compliance enforcement
The Backend Infrastructure
- FastAPI Python backend with PostgreSQL and read replicas
- Multi-tenant architecture with vendor isolation at the database level
- Auth0 JWT authentication with role-based access control
- Celery + Redis for background task processing
- Alembic for zero-downtime database migrations
Compliance, Built In From Day One
This is where my background really mattered. I didn't add compliance as a feature — the entire platform is built around it:
- Automated Certificate of Analysis (COA) management with QR code tracking
- Age verification that produces verifiable records, not just checkboxes
- Inspection-ready Audit Packs: signed, verifiable records on demand
- License validation and automated expiration tracking
- A regulatory rules engine designed so new jurisdictions are configuration, not rewrites
- Audit trails for every transaction, every delivery, every compliance checkpoint
Scaling With AI Instead of Headcount
Here's where the machine learning background became a force multiplier. We run a 7-person AI-augmented team, and AI agents carry a real share of the operational load:
- An automated content pipeline with a zero-tolerance compliance gate — it rejects its own output when the copy doesn't meet regulatory standards
- Agent-driven operations automation for the work that usually demands whole departments
- Reliability patterns — circuit breakers, idempotent jobs, structured logging — that let a small team run like a much bigger one
Why Hemp? Why Regulated Commerce?
This question comes up in almost every conversation. The answer is multifaceted.
First, it's the perfect application of my skillset. Regulatory compliance, operations, inventory, analytics — these are software problems that reward engineering depth. Add machine learning, and you've got a technical challenge worth a decade of attention.
Second, the timing is real. Texas hemp retailers are now operating under 25 TAC Chapter 300 with DSHS inspections active. The businesses that get audit-ready will be the ones that last, and most of them don't have the tools yet.
Third, hard markets reward the builders who show up. The more demanding the regulatory environment, the more an operator needs software that was designed for it. We built for Texas — arguably the most challenging state — and designed the platform so that other jurisdictions are configuration, not rewrites.
Finally, it aligns with my values. Hemp retailers are legitimate business owners serving customers who want safe, verified products. Those benefits are meaningless if the products aren't compliant and the records can't be trusted. I'm building the infrastructure that makes trustworthy commerce possible at scale.
The Mission: Helping Hemp Retailers Grow With Confidence
HempDash is the AI Operating System for the Hemp Industry, and every word of that matters.
"AI Operating System" — Not a point tool. Retailers start with compliance and expand into operations: COA management, age verification, Audit Packs, delivery, analytics, and AI-assisted workflows in one platform.
"For the Hemp Industry" — Built for hemp (<0.3% THC) businesses operating under real regulations, starting with Texas and 25 TAC Chapter 300. Compliance isn't a checkbox here; it's the foundation.
"Grow with confidence" — The point isn't paperwork. It's protecting the business, saving time, and reducing risk so owners can focus on growing revenue.
Why This Matters: Beyond the Business
I could have stayed in Silicon Valley building the next AI chatbot. Those are important problems, but they're crowded problems with hundreds of well-funded teams attacking them.
The hemp industry needs technical sophistication. It needs someone who can navigate regulatory complexity with the same rigor we apply to distributed systems architecture. It needs infrastructure that can scale beyond Texas without breaking.
Most importantly, it needs someone who understands that compliance isn't a burden — it's the enabling technology that makes everything else possible.
I'm building HempDash because regulated industries deserve the same technical excellence that unregulated industries take for granted. Retailers deserve infrastructure that protects them instead of exposing them to risk. And their customers deserve safe, verified products.
That's why I left Google. And that's why I spent 14 months building before launch.