Alasdair Allan is a scientist, author, hacker, and journalist. He is the CTO of Negroni Venture Studios, a Cambridge-based seed and angel-level VC firm backing embedded hardware and vertically integrated companies.

He works at the intersection of open hardware, machine learning, and embedded systems — particularly wireless sensor networks. As part of his role at Negroni he recently spent a year as interim CTO at portfolio company Evaro, leading the engineering team through a major replatforming effort.

Before Negroni he was Head of Documentation at Raspberry Pi, where he led the team responsible for everything from beginner-friendly tutorials to register-level documentation of new silicon. He was, in other words, the person responsible for writing things down.

He has written eight books, over a hundred academic papers, and several standards covering real-time events and application interoperability. You can often find him at conferences talking about interesting things, or deploying sensors to measure them. Sometimes he also stands in front of cameras.

Alasdair is a recovering astrophysicist. He built a peer-to-peer network of telescopes that autonomously scheduled observations of time-critical events, contributing to the detection of what was, at the time, the most distant object yet discovered.

He's known for benchmarking the new generation of Tiny ML accelerator hardware and, somewhat inexplicably, hacking hotel radios.

Some years ago he rolled out a mesh network of five hundred sensor motes covering the whole of Moscone West during Google I/O. A couple of years before that he caused a privacy scandal by uncovering that your iPhone was recording your location all the time. One of the first big scandals around mobile data, it later became known as "locationgate" and triggered several class action lawsuits and a U.S. Senate hearing. It even got a mention on South Park. Some years on, he still isn't sure what to think about that.

In the past he has written for Make: Magazine, Motherboard and VICE, Hackaday, Hackster.io, Raspberry Pi, and the O'Reilly Radar.

He's currently building Vera, a programming language designed for machines to write rather than humans. It throws out variable names entirely, requires contracts on every function, and compiles to WebAssembly. The idea: if LLMs are going to write most of the code, the language should be designed around what they're good at and bad at, rather than pretending they're people.