Integrated operating system for the solar industry
Industry
Energy
Use Case
Autonomous Solar Mapping
Company size
100 employees
Headquarters
Somerville, Massachusetts
Data labelling
Discover our solutions
About
Raptor Maps is a Series B scale-up on a mission to build a more scalable and resilient solar industry.
Solar is one of the top renewable energy resources driving the clean energy transformation, but the sector is facing critical challenges. Underperforming power plants, labor constraints, tightening margins, and supply chain issues are just a few of the problems negatively impacting scalability, efficiency, and revenue.
The team behind Raptor Maps realized that most of these challenges could be solved with software. That’s why they developed a monitoring, inspection, and remediation platform, providing digital tools to identify solar sites’ issues and inefficiencies.
Raptor Maps offers an integrated operating system that enables end-to-end solar management, drone-in-a-box, and other robotics use cases across the solar project lifecycle, as well as comprehensive photovoltaic (PV) inspections. With the help of robotics and AI, their platform can aggregate different types of visual data, including visible spectrum and thermal infrared imagery, and identify potential risk points.
The challenge
Raptor Maps builds software for the solar industry that uses actionable digital tools to identify and address underperformance in solar sites.
The team integrates insights from robotics, drones, and ground robots, collecting invaluable data to fuel operational and remediation workflows.
In order to speed up R&D and deliver solutions to their customers faster, they began searching for a training data platform that would enable rapid annotation of large datasets and human-in-the-loop management, replacing their previous outdated in-house data labeling tool, which no longer met industry standards.
The solution
With the help of V7’s data annotation toolkit, automated workflows, and labeling services, Raptor Maps reduced their time-to-model from approximately one year to 4-8 weeks.
The V7 platform easily handles vast data volumes and provides annotation and project management features that allow the team to significantly streamline their product development, as well as reduce operational costs. The labeling workforce took over most of the labeling tasks from the in-house team, improving their capacity.
The Raptor Maps team can now deliver solutions to their customers faster and more cost-efficiently, helping them advance their research and support the renewable energy sector.
The results
Raptor Maps has already reduced time-to-model by 83%, and their goal is to streamline the model iteration loop even further.
With V7’s help, the Raptor Maps team can get prototypes to customers much faster and test them on the job—which lets them try out more solutions, explore more use cases, and advance continuous robotics in the renewable energy industry. Additionally, this significant improvement in delivery time helps build better customer relationships, minimize costs, and eliminate more hazards in less time.
Raptor Maps is now well-equipped to face all the challenges posed by the rapid evolution of the solar industry. With the growth of solar farms and continuous robotics, the data outputs from the plants increase—the deployed drones can now deliver a high-quality image every two seconds. By refining their technology, the Raptor Maps team wants to be able to process and analyze large amounts of data near the edge.
The team is also excited to keep advancing continuous robotics solutions—with the goal of bringing the industry one step closer to the reality in which robots work, inspect, and maintain solar farms autonomously.
V7 is proud to support Raptor Maps on its mission to introduce continuous robotics in solar plants globally and help the solar industry overcome the obstacles even faster—letting the renewable energy sector thrive and meet global climate goals.
Training data needs
In 2015, when Raptor Maps was only beginning its machine-learning journey, the training data platform market was in its infancy. With no solutions catering to specialized use cases to choose from, the team was forced to build its own internal data annotation tool.
Five years later, the in-house software was no longer meeting market demands. The team realized it was outdated and unfit for more advanced technologies used by their customers. However, rebuilding and maintaining the tool would drain lots of resources. At the same time, the training data platform market has become much more robust, with multiple tools able to handle even the most edge cases.
Faced with the build vs. buy dilemma, they decided to search for a partner to access more up-to-date technology, ensure the highest scalability, and focus in-house engineering talent on other strategic projects.
When designing solutions for solar industry customers, the Raptor Map team often starts with a proof of concept— preparing a prototype and using customer feedback to iterate and improve. For each new use case, the team must gather data to annotate and train a machine learning model.
This engineering loop could last up to a year, causing bottlenecks and massively extending delivery times. In an era of out-of-the-box solutions, Raptor Map knew they needed to streamline their process to satisfy their customers.
After an exhaustive evaluation of multiple solutions, the team decided to go with V7—due to its ability to go far beyond traditional data labeling capabilities.
Testimonial
Eddie Obropta
Co-founder & CTO of Raptor Maps
Why V7?
Working with V7 helped the Raptor Maps team drastically improve their time-to-model—going from 12 months to 4-8 weeks on average.
The Raptor Maps team uses V7’s auto-annotation and model-in-the-loop features to annotate data much faster than their previous internal solution allowed them to. They also take full advantage of the flexibility of the annotation toolkit. The team can now work with data in different stages of the annotation process—labeling data from scratch, improving existing annotations, or using ML models for pre-annotation.
Another V7 feature that turned out to be a game-changer was the transparent, drag-and-drop workflows. Previously, project management posed a huge organizational challenge due to the complexities of solar industry projects. With V7, the Raptor Maps team can easily manage even the most complicated pipelines, monitor human-in-the-loop processes in real time, tighten feedback and quality assurance cycles, and track statistics to benchmark and improve.
Raptor Maps also cooperates with V7’s labeling workforce. Previously, the in-house team took care of all annotations, which reduced their capacity to work on solar-industry-related tasks. Now, the internal team only reviews the external labelers’ annotations, drastically speeding up delivery times.
On top of that, the Raptor Maps team values V7’s dedication to data security—including GDPR compliance and a strong understanding of AI adopters’ security needs—which lets them safely train and deploy their models.