Ross Melbourne

Innovations

Building things is in my family’s DNA.

Some of it shipped to thousands of companies. Some of it was a first of its kind. All of it started with the same question — why does this not exist yet?

01Physical AIActive
Yuri, a dual-arm humanoid robot on its mobile base
Fig. 01 — Yuri, mobile baseSouthlake, TX

Sensori Robotics

I co-founded Sensori Robotics with my brother David to do the thing I had wanted to work on my whole career.

Sensori builds Yuri, a dual-arm humanoid platform for teams training vision-language-action models. The problem it solves is unglamorous and expensive: labs doing serious Physical AI research lose their first months to integration — sourcing arms, wiring sensors, fighting drivers, keeping a bench rig alive — before a single useful frame of data gets collected.

Yuri ships as a complete machine instead. Everything a lab needs to start collecting data on the first day arrives in one crate, and a mobile base is available if the work needs to leave the bench. We build it in Southlake, Texas.

sensorirobotics.com ↗

Yuri in simulation
Fig. 02 — Yuri in simulation
02Applied AI and writing
Moral Code, a novel by Lois Melbourne
Fig. 03 — Moral CodeA novel by Lois Melbourne

The book

Moral Code

Lois wrote a novel, and I helped. An ethical AI merges with nanite robots and sets out to protect every child on earth from serious harm.

The idea came out of a question I kept returning to. What happens when computers get small enough to be as light as dust? What if that dust could assemble itself into machines of any size, shaped to any job? And what if the intelligence directing it were advanced enough not only to think like a person, but to feel the protective instinct a parent feels?

All proceeds go to THORN, an AI-powered organisation building a safer world for all kids — which is close to the thing the novel imagines.

moralcodethebook.com ↗

A conversation with EllyBot
Fig. 04 — EllyBotMid-2022

Applied AI · Retired

EllyBot

Elly is the AI character at the centre of Moral Code. I built a version of her you could actually talk to.

I built EllyBot in the middle of 2022, on OpenAI’s GPT-3 API. ChatGPT was released that November — so this shipped roughly half a year before the thing that made everyone else start paying attention.

At the time there was no category for it. No chat products, no prompt engineering, no expectation that a novel might come with a conversational character attached. It was simply the most interesting thing available to build with, so I built it.

EllyBot is no longer running. The idea it was testing — that a character could be a piece of software you have a relationship with — turned out to be most of what the next three years were about.

03Enterprise software1989 — 2011
An organisation chart drawn automatically by OrgPublisher
Fig. 05 — OrgPublisherFirst shipped 1995

OrgPublisher

In 1989 I started coding the first version of OrgPublisher. It drew a company’s organisation chart automatically, from the personnel data that company already had. It took six years to become a product: the first copy sold and shipped in 1995.

Before that, org charts were drawn by hand and out of date the day they were finished. OrgPublisher made them a by-product of data you were already keeping. Automatic org charting is now an industry standard, and this was the first of them.

It became Aquire, the enterprise software company I co-founded with my wife Lois. We grew it to over 3,000 customers in more than 100 countries and sold it in 2011 with zero debt and 100% ownership.

3,000+
Customers
100+
Countries
2011
Sold, zero debt
HR product award seal

InSight

An award-winning organisational analytics product that turned org chart history into workforce intelligence — promotions, transfers, terminations, new hires, and how talent actually moved through a company over time. The patent below came out of this work.

The insight was that the charts were already the record — years of who-reported-to-whom, tracked over time, already described how careers moved through an organisation. InSight made that record legible.

InSight attrition analytics
Fig. 06 — InSight analytics
InSight talent pipeline view
Fig. 07 — Talent pipeline

Patents

Six patents in total — four granted, two pending. Two cover workforce analytics, including US 12/893,890 for automatic career-movement analysis. The other four cover the robotics work at Sensori.

04GamesFirst of its kind

Save Dallas

Before Save Dallas, game worlds were hand-crafted to resemble a real place or invent a fantasy one. Save Dallas used an actual 3D model of downtown Dallas, and it ran on a phone — iPhone and Android.

As far as we could establish, it was the first mobile game built on a real city model.

Save Dallas — a fighter and mothership over downtown Dallas
Fig. 08 — Save Dallas, iOS
Save Dallas — mothership hovering over the Dallas skyline
Fig. 09 — Save Dallas, iOS