Amanda Siqueira

Amanda Siqueira got into water through the single most tedious job the sector has to offer. Early in her career, as an engineering intern, she spent eight hours a day watching CCTV footage from inside sewer pipes, writing down every crack, root and blockage she could spot. Her own line for it on the show was blunt: without software, that whole process is "painfully manual," the kind of thing that turns weeks of an engineer's life into a backlog. So when the open-source deep-learning frameworks (the toolkits that let computers learn to recognise things in images) became usable around 2017, she and her school-and-university friend Michelle Aguilar, a mechatronics engineer, looked at that footage and saw a problem a machine could finally take over.

Amanda Siqueira and Michelle Aguilar co-founded VAPAR in 2018, and the clever part is the moat they picked. Reading the inside of a sewer pipe is a rare skill, and as Siqueira put it, there are only "some 3,000, some 5,000 people globally that know what they're looking at when they're looking at the inside of a sewer pipe." That scarcity, which sounds like a reason to avoid the problem, was exactly why it was defensible: a generalist tech giant would find it very hard to replicate the domain expertise needed to train the models. So VAPAR went narrow on purpose, building AI that codes in-pipe inspection video automatically and then layers risk and cost on top of the defects, so a utility can answer the only two questions that matter: what do I fix, and how much will it cost.

VAPAR exists because the economics of pipes are brutal. Putting a camera in a pipe is cheap, somewhere between $5 and $10 a metre in Australia, but actually rehabilitating that pipe runs from $250 to $1,000 a metre, so the expensive mistake is spending rehab money on the wrong asset. Most utilities still run reactively, waiting until a customer calls about sewage in the basement before they send a camera down, and globally over a billion metres of pipe is inspected every year, with every metre often watched twice because a manual review is error-prone. VAPAR's whole pitch is to flip that around: read all the footage with AI, rank the assets by real condition, and fix pipes before they fail rather than after the sinkhole.

Amanda Siqueira has taken that thesis from a single Sydney council, which handed VAPAR 5,000 uncoded inspection videos as its very first project, to utilities on three continents. She points to the UK as VAPAR's biggest market, where Northumbrian Water now uploads its reactive CCTV footage straight to the platform and lets the AI do the coding before its own specialists sign off, a model she calls keeping the expert in the loop. With about $5.6 million raised across a 2020 seed, a 2022 seed and a 2024 Series A led by PureTerra Ventures (with Autodesk also writing a cheque), the company is pushing into the United States while consolidating Britain. The ambition, in her words, runs "much, much higher than just doing pipes," toward the wider base of public infrastructure that gets maintained with taxpayers' money.

“I thought that I needed to get an MBA. So I spent about a year trying to get into an MBA, and another engineering colleague of mine who had started his own business said, you just got to start. There's no MBA in the world that's going to teach you anything.”

She is, in short, the founder who did the boring job first and then built the machine to kill it, which is most of why VAPAR reads like a company solving a problem it actually understands.