In the past year we participated in a number of virtual hackathons and won ✨ five awards ✨ from Google AI, WISH, COVIDathon and other initiatives supported by health, technology and government organizations.

We rolled out a new People 🍃 Fall Detection model, to help families and caregivers care remotely for their loved ones isolated at home during the COVID lockdown. See a short promo video about our Fall Detector.

We also introduced the 🔋 Ambianic Box project which makes it very easy to assemble and install a new Ambianic Edge device without any programming skills.

A few other notable achievements in 2020:

  • Always-on timeline. Access your Ambianic Box at any time from anywhere.
  • Over-the-air updates. Automatically deployed by our CI/CD system with each release.
  • Installable UI app for mobile and desktop at
  • Instant notifications: email, SMS, Slack.

We have also grown as a 🌎 community. There are over 700 👨‍👩‍👧‍👦 members of our public Slack channel.

I am also very excited that our team has expanded with 12 new code contributors!

The looming trend 🌊 of smarts surveillance camera 📹 vendors failing to protect user privacy continued with 🚩 🚩 new reports of major breaches 🚩 🚩 across not only individual consumer cameras but also powerful and very tech savvy enterprises such as 🚗 Tesla and ☁️ Cloudflare as well as government organizations, jails, schools, police stations and 🏥 hospitals.

Our conviction is stronger than ever that protecting user privacy is essential to win the trust of end users and remove the stigma created by big tech. AI 🧠 can do a lot of good for many people if its implemented in a transparent and ethical way. No foggy 🌁 promises and obscure fine print.

It has been greatly motivating to see the grass roots adoption of Federated Learning. Two years ago there was only one Open Source FL framework available - Tensorflow Federated. Now there are at least 9!

Along with the growth of Federated Learning we are starting to see AI luminaries come out and speak openly about the unscrupulous attempts of big tech to control AI adoption. More than that, we are seeing that Open Source AI models are becoming good enough for many practical applications as long as there is good data available to power these models.

All these developments and the great feedback we received from users helped us shape the 2021 roadmap. Here are the big project milestones for 2021:

  • Easy access, monitoring and management of multiple edge devices via UI app.
  • Live view - Heads Up Display (HUD) in UI app. Show in near-real time what source samples are coming in along with AI inferencing info.
  • Enable users to label timeline events. Use feedback to incrementally learn AI models and improve accuracy. For example label detected faces to train local on-device recognition.
  • Local model training on a single or multiple devices owned by the same user. For example face recognition of family or employees.
  • Federated model training across devices owned by many users. For example ALRP (auto license plate recognition).
  • Visual edge device config and pipeline flow design.

We look forward ⏩ to share together in 2021 the rebirth of Ambient Intelligence: build for the people and by the people. We will continue to demonstrate that it is possible to empower users to pool their resources and create helpful AI applications without giving up any control and privacy.

Stay healthy and thank you for your continued support! 🙏

Ivelin Ivanov Founder,