Expertise
Decades of building software where the stakes are high and the requirements are unusual, across industries that rarely overlap and a wide range of languages and platforms.
Areas of concentration
AI, speech, and video generation
We build with modern AI: language models grounded in private knowledge bases, so answers come from your documents instead of blindly relying on the public internet; speech recognition and synthesis, so people can talk to a system and hear it answer; and generated video. The hard part is rarely the model itself. It’s the plumbing around it: getting the right data in, keeping answers accurate, holding latency and cost down, and fitting it into the systems and workflows people already use.
Levyco AI, coming to this site, is a working example: ask it a question, by typing or by voice, and it answers from our own knowledge base.
Worldwide public gaming and lotteries
Public lotteries run on software that cannot be wrong. Every ticket has to be accounted for, every draw has to be provably fair, and every transaction has to survive an audit by a regulator. We’ve built systems for lotteries operating around the world, where traffic spikes in the minutes before a draw, retailers depend on terminals that simply work, and the rules change from one jurisdiction to the next.
Online gaming and skill games
Online games of skill combine real-time play with real money, which means fairness, security, and speed all at once. We’ve built game engines, matchmaking and tournaments, wallets and payouts, and the controls that keep play honest: anti-cheat measures, geolocation and age checks, and the record-keeping regulators expect.
Fintech
In financial software, a rounding error is a defect and a missing record is an incident. We build payment flows, ledgers, reconciliation, and integrations with banks and processors, designed so that every cent can be traced and every operation can be repeated safely when a network fails halfway through.
Retail
Retail is many businesses under one roof, and we’ve built software for most of them: sales and point of sale, purchasing, technical support, inventory, warehousing and shipping, repairs, and pricing. The value is usually in connecting them, so that a sale updates stock, stock drives purchasing, a repair is tracked from intake to return, and prices change everywhere at once.
Manufacturing and printing
Manufacturing software lives where data meets machinery. We’ve built production scheduling, print workflows and variable-data printing, machine integration, and quality control: systems that turn orders into physical output, on time and to specification, and know when something on the floor has gone wrong.
Game show technology
Television is live, and live means there’s no second take. We’ve built the technology behind game shows: contestant systems, scoring and timing, displays and on-screen graphics, and the control software that ties them together, designed to work the first time, every time, with a studio audience watching.
Research and development
Some projects start with a question rather than a specification: can this be done, how fast can it be, what would it take? We do focused research and prototyping to answer those questions quickly and honestly, so that the decision to build, or not, is based on evidence.
Technology expertise
Languages
We work across the stack, in the language that fits the problem. For systems and performance-critical work: C++ (including C and Objective-C) and assembler, including low-level GPU programming, such as fragment and compute shaders. For graphics and real-time rendering: Metal and DirectX. For servers, services, and the web: JavaScript and Node.js, Java, Go, PHP, and Python. For mobile apps: Swift and Kotlin.
Breadth matters because real systems are mixtures. A single project might pair a C++ engine with a Node.js service, a Python data pipeline, and native apps on two phone platforms, and the seams between them are where problems hide.
Platforms
We’ve shipped software on desktops (Windows, macOS, Linux, and Solaris), on mobile devices, on embedded systems, and on custom hardware built for a single purpose. Each brings its own constraints: memory and power on a device, certification on specialized hardware, distribution and updates on desktops and phones.
Infrastructure and automation
Building infrastructure and automating deployments of insfrastructure and code in cloud platforms has become a crtitical skillset that we use everyday.
Not limited to a list
The languages and platforms above are where we’ve spent the most time, not the edge of what we’ll do. We’re eager and happy to learn and adapt to any environment, and a long history across so many industries and technologies makes learning a new one fast and effective. If your project runs on something not listed here, that’s a conversation we’d like to have.