Can a local LLM answer as fast as the cloud?
Benchmarking full voice pipelines on consumer GPUs to see whether a private, local assistant can match the responsiveness of cloud services.
Black Sheep Group Independent technology lab
We build things differently. Independent software, AI systems, developer tools, research, and experiments: dreamed up, tested properly, and improved by what we learn.
4 projects. Different problems, one way of working.
AI / Smart Home / Voice
Whole-home AI that listens, understands, and responds wherever you are.
Finance / Research / Data
Tools for exploring markets, investments, and the data behind them.
Bible Study / Education / Digital Resources
Bible study resources built to make serious study easier to access.
Developer Tools / Databases
A modern SQL workspace that lives in your browser.
The projects share no market. They share a method.
Start with what the software could be, not how it has always worked. Most of our projects begin with "what if it didn't have to be like this?"
Build it and measure it: benchmarks, prototypes, pilots with real users. Until it's been tested, an idea is still a guess.
Keep what works, drop what doesn't, and write down why. What we learn goes into the next version, or the next project.
Not every project goes through every stage. Experiments start in Labs.
Benchmarks, experiments, and what we learned from them.
Benchmarking full voice pipelines on consumer GPUs to see whether a private, local assistant can match the responsiveness of cloud services.
Testing whether wake-word loudness can pick the nearest device in a multi-room setup built from commodity speakerphones.