Video pitch

Video demo


Currently, the world’s cities generate about 1.3 billion tonnes of municipal solid waste (MSW) per year and the amount is expected to double by 2025 due to rapid urbanization, mass consumption and throw-away lifestyles (Gardner, 2012). Rapid urbanization, population growth, migration to urban areas, lack of sufficient funds, and affordable services often force city authorities to offer unreliable and inefficient waste management services (Wilson et al., 2006). In many developing countries, city authorities often collect only 50 to 80% of waste and open dumping and landfills are frequently the only available disposal options (Medina and Dows, 2000).

Today, many developed cities such as Stockholm, and Adelaide are aiming to transform their current waste management practice into a more efficient and sustainable way. (Zaman and Lehmann,2011). Zero waste is a visionary concept for confronting waste problems in our society. The idea is being developed and implemented in various sectors including waste management and treatment, mining, manufacturing, and urban development. The zero-waste concept has been embraced by policymakers because it stimulates sustainable production and consumption, optimum recycling, and resource recovery. (Zaman, 2015)

However, transforming currently over-consuming cities into zero-waste cities is challenging. Therefore, we accept the challenge of providing emerging technologies to transform cities and citizens to implement the zero-waste concept in their lives.

Hoornweg, Daniel; Bhada-Tata, Perinaz. 2012. What a Waste: A Global Review of Solid Waste Management. Urban development series; knowledge papers no. 15. World Bank, Washington, DC. © World Bank. License: CC BY 3.0 IGO.”

Zaman, A. U. (2015). A comprehensive review of the development of zero waste management: lessons learned and guidelines. Journal of Cleaner Production, 91, 12-25.

Zaman, A. U., & Lehmann, S. (2013). The zero-waste index: a performance measurement tool for waste management systems in a ‘zero waste city’. Journal of cleaner production, 50, 123-132.

Zaman, A. U., & Lehmann, S. (2011). Urban growth and waste management optimization towards ‘zero waste city’. City, Culture and Society, 2(4), 177-187.

What it does

Debrazio’s purpose is to use Zero Waste Management, a holistic waste management a concept that recognizes waste both as a resource and a symbol of the inefficiency of our modern society (Zaman and Lehmann, 2013).

Debris + Iguazio = Debrazio. In this sense, our goal is to develop a tool based on emerging technologies such as the Internet of Things (IoT) and Machine Learning (ML) to improve waste management.

Our application will help our customers to apply new emerging technologies for waste management, focusing on the circular economy of raw materials, recovery waste energy, composting, and contributing to the protection of the environment. Basically, we are going to provide an application to detect types of waste and turn it into something profitable and we have more plans of working on waste management facilities using IoT.

How we built it

This is a team effort and everyone contributed one way or another through conceptualization and software testing, and development.

Phase 1: Find members - We came together by searching and actively looking for members in the Slack group, MLOps Live.

Phase 2: Brainstorming - A lot of discussions happened since June 16. It includes choosing which problem to solve, two of which were (Idea#1) Strategic Social Services Building and (Idea#2) Global Reporting Initiative standards (GRI). We kept in mind how we would sell the idea to potential investors. We also considered if there will be publicly available datasets and how feasible each idea would be.

Phase 3: Planning: Tasks were assigned for each member from the conceptualization, MLRun backend, and frontend development. For the dataset, we also found public repositories and have considered , as prescribed by Gilad. The Research part is spread out through the time period and is still ongoing.

Phase 4: Development Early stages include running Iguazio free cluster provided (Thanks to the Azure access provided by Gilad). Once we are able to use Iguazio platform, our task was to start MLRun using the Jupyter notebooks provided getting started documentation. We followed the four (4) main steps to successfully drive the project: running a function on a cluster, training a model, deploying in service, and creating a pipeline. It took a while for the code to run and to understand some concepts for given lines of Python code but soon it became easier. There were also a lot of times when we had to consult Gilad and he gladly helped the team progress until we are able to create a pipeline and be able to view its flow diagram in the Iguazio dashboard.

Phase 5: Preparing submission Ongoing

Challenges we ran into

Initially, none of the team members have experience with Iguazio and MLRun. Moreover, very few have used python notebooks so it took more than a few weeks to familiarize ourselves with the stack. There were also blockers in using Iguazio's pre-made code but Gilad was able to help the team proceed.

Accomplishments that we're proud of

The team of five (5) members from three (3) different time zones were able to build a project that has an impact on the environment in a very limited period of time. But we are not done.

What we learned

For beginners, we're happy to familiarize ourselves with the use of Iguazio Platform, how easy it is to set up and get things started very quickly.


Our product innovation goes beyond using the Internet of Things and Machine Learning, we want to make waste profitable to encourage even more citizens to choose a lifestyle based on recycling and composting.


Our products and services will be a solution possibly embraced by environmental organizations, sustainable companies (Business to business - B2B), customers (Business to customer - B2C), and the government (Business to government - B2G). See the Business Canvas in our gallery


Starting with our app, our goal is to make our tool available to the citizens of Brazil, as the country's waste management is a concern, as you can see in this news about overseas waste pollution on beaches in Northeast Brazil.

With more than 200 million inhabitants, Brazil is one of the countries that generate the most solid waste - materials, substances, and discarded objects - whose final destination should be treated with economically viable solutions, according to the legislation and technologies currently available, but they end up even in part, being dumped in the open, thrown into the public sewer system or even burned.

Thinking about solutions to the problem, we thought about how we could take advantage of waste to generate income for people, since in Brazil, according to a study, there are many collectors in Brazil who survive from the sale of waste and are responsible for more than 90% of recycled materials in the country.

What's next for Debrazio

Our roadmap includes:

  • Planning waste management facilities using sensors / IoT - Business to business (B2B) and Business to government (B2G) (See this link to know more)
  • Detect hazardous and medical waste
  • Build a sample waste container (see uploaded image labeled as Glass / Plastic / Metal Sample Classifier) and integrate with IoT (Machine vision embedded system, please see uploaded image) and MLRun backend

Datasets considered

Google marketplace

AWS Open Data

Azure Open Datasets


Kaggle Dogs Vs Cats

Kaggle Waste Classification

Project repositories



Our team

Name / Country / Course / Profession:

  1. Eugene De Los Santos / Philippines / Electronics / SW Developer / Linkedin
  2. Daryl
  3. Gabriella / Brazil / Biological Sciences / Frontend Developer
  4. Anderson / Brazil / Accounting / IoT Prototype
+ 13 more
Share this project: