A summary of Zabble's AI object detection models development process and how we use industry metrics and benchmarks to develop, test and evaluate our AI systems.
Zabble's AI Engineer, Alberto De La Rosa, and Customer Success Manager, Derek Davis, met up for the first time and went full San Francisco when they rode in a driverless rideshare car operated by Waymo in the city. They recall that the absence of a visible driver felt strange, especially when the car starts navigating traffic on its own. Impressed with how well the car handles itself but anxious during sudden braking or sharp lane changes, Alberto and Derek were fascinated by the car. Perhaps this is a shared experience for the use of any new technology.
They were on a short trip to the San Francisco Bay Area to visit some of Zabble’s key partners, PSSI and Stanford, UCSF, UCSF Health and the Golden State Warriors to learn about how they use Zabble’s products to track progress towards zero waste. In this blog, we are going to share a couple of different experiences as witnessed by Derek and Alberto who were later joined by other team members Nik Balachandran and Kelley Nelson.
Alberto usually finds himself planted in front of a five monitor set-up, some of which are flooded with windows of code and one reserved for your typical work-from-home essentials like Zoom and Slack. Today, he isn’t on a conference call or writing complex code. Instead, Alberto is cruising in the passenger seat of a garbage truck, ready to look at some trash! He was accompanied by Javier, the recycling truck driver, on the ride-along, where they bonded over their shared hispanic heritage, as Javier eagerly learned more about how Zabble’s AI worked. He learned how important it was for the driver to service the routes efficiently and stay on time. Oftentimes, Javier’s routes would change owing to seasonal variation in their campus client’s activities or frequent right-sizing of service levels. Today, he had to pay attention to which stops he would be serving on his route. Luckily, Katherine, from the office, had made updates the previous evening and when he opened his Zabble Mobile Tagging app this morning, he was able to view his stops on a map. “It’s great to not have to worry about internet connection when using this app because there are some patchy spots on campus”, he says smiling.
Sharing this cold morning on a different truck is Derek Davis, Zabble’s burgeoning customer success manager who traveled from Florida to experience first-hand what Zabble’s customers go through every day as they collect materials for their campuses. Derek was accompanied by Albert, the landfill truck driver, who was candid about some of his frustrations around learning/using new technologies like truck cameras and AI mobile apps, but conceded that they were helpful and important for campus operations because it gives the university the visibility it needs to make informed decisions to reduce their waste footprint. Derek also quickly realized that the ride-along involved more than just sitting in the truck and pushing buttons as there was a lot of getting off, pushing heavy bins around and then moving on to the next. He noticed that most of the rolling bins were locked or tucked away in enclosures, thus lengthening the collection process. Sometimes these enclosures are blocked preventing the bins from being serviced or there are bulky items on the ground. Other times, they are contaminated. “I see a lot of cardboard in these (landfill) boxes”, says Albert as he empties a front-end loader bin into the hopper of his truck. All these cases have to be communicated to the customer (“building manager”) so they can help prevent this from happening again. Before the PSSI team used Zabble, handwritten notes were shared with dispatch at the end of the day which were then shared with their customer. Since they have been using Zabble’s Mobile Tagging platform, they have been able to automatically send email notifications to their university customer. “To reduce contamination, we have to keep educating them!”, he concludes.
Unmistakably, the drivers took their jobs very seriously and were enthusiastic to share every day experiences with Derek and Alberto. Both the drivers took immense pride in their duties, were curious about AI and technology and reminded Derek and Alberto that it was important for them to see how this data was being used by their client. The Zabble team greatly appreciated this opportunity to learn more from the PSSI team to understand their everyday challenges in order to augment their productivity and drive more value for its clients.
This time Derek and Alberto are accompanied by Kelley Nelson, Director of Product and CEO Nikhil Balachandran. The UCSF campus is in a more urban location - more compact and vertical - calling for a different use-case than the previous campus site visit. We followed Jon, an experienced waste reduction and recycling supervisor, who consolidates materials from dumpsters in nearby buildings on his daily routes into a compactor. The loading docks at some of these buildings are prime grounds where the materials are hand-sorted and transferred into compactors. Ben and Jermaine are in charge of sorting today. Insights on specific contamination issues are shared with the building managers and custodial supervisors to help close the loop between observations and action. “I really want someone to look at the data and analytics and do something with it," says Ben, as pulls up another bag from the landfill dumpster to sort. But the internet here is spotty. This manual sifting calls for a technology solution that can produce information at their fingertips, even without needing access to the internet, uninterrupted. The UCSF team has grown to eight members now from three in 2019, making it one of Zabble’s oldest clients. CEO Nik Balachandran has made these onsite visits a priority for the company in order to assess exactly what improvements in the product our clients would benefit from. Nik recognized the struggle in this experience several years ago, when he started Zabble. Much before the AI boom, he saw the power in machine learning and AI to capture waste data for easier real-time reporting and action. He always makes it a point to do these trips because he says that we have to get our hands “dirty” in order to really solve the problem.
For Kelley, who has been with Zabble since 2020, this isn’t just about the zero waste goals or the landfill diversion improvements that UCSF’s Daniel Chau, Assistant Director of Recycling and Waste Reduction Program, can happily boast about (they currently have the best diversion rate in the state for a public university**) but showing the customer that they are fully invested, dedicated to the cause of zero waste. When she realized that Zabble’s clients needed to communicate with each other about contamination and other incidents internally, she spearheaded the effort to build a Zabble feature called Alerts that automatically sent real-time email notifications with the location, photo and timestamp of the incident to a pre-assigned list of stakeholders. This feature has come in handy for the UCSF team when communicating the presence of hazardous items in the bags to their EH&S teams in real-time to act upon immediately.
The team then got a quick tour of the labs to see how supervisor Glen uses the Zabble platform to keep the lab plastic contamination rate under 5% before it gets picked up by a third-party vendor. “Constant monitoring and communication is key to making this program successful”, he says, proud of what he has accomplished in a short span of time. Not long ago UCSF was only operating the lab plastics program in a handful of buildings. Now, they are in the midst of scaling this program across all of their three campuses in San Francisco.
The value of the firsthand experience gained from visiting a customer site is immeasurable. The team feels energized and rejuvenated after the culmination of these enriching client visits. The discussion, feedback and observations call for the introduction of Zabble 3.0, the next era of AI that goes beyond identification of materials in the bins or bags. It is the cumulative effect of taking action to report incidents in real-time based on the distilling of granular and meaningful insights and proactive engagement with staff and occupants to close the loop, constantly.
It takes collective effort from everyone to achieve zero waste!
Reach out to us and ask us how we can help get you to zero waste!
This blog post discusses Zabble's team visits to key partners, PSSI, Stanford and UCSF, to observe a day in their lives and how they use Zabble's products for zero waste initiatives. At PSSI, Zabble's team rode along with garbage and recycling truck drivers, observing challenges like route changes and contamination. They learned about the importance of real-time communication and notifications for addressing these issues. At UCSF, the team witnessed manual waste sorting and the need for technology that works offline. They also saw how Zabble's "Alerts" feature helps communicate contamination and hazardous items. The blog concludes with the value of these visits and the vision for Zabble 3.0, which aims to provide more granular insights and proactive engagement for achieving zero waste.
Wednesday, April 23, 2025
A summary of Zabble's AI object detection models development process and how we use industry metrics and benchmarks to develop, test and evaluate our AI systems.
Wednesday, April 2, 2025
Read about the basics of object detection AI, challenges to developing a methodology and metrics and what it means for contamination detection and identification at the bin.
Monday, March 31, 2025
1966 Tice Valley Blvd, #105,
Walnut Creek, CA 94595
Tel.: 925-289-9345
Email: team@zabbleinc.com
Mobile Tagging
U.S. Patent No. 11,788,877