The infrastructure running Humen.Ai’s popular iOS app became burdensome to run and threatened its financial stability. The company needed to lay a foundation for rapid releases and quickly reduce infrastructure costs without impacting response times.Lösungen Cloud, Public Cloud, Data, Artificial Intelligence & Machine Learning Plattformen Cloud, Amazon Web Services (AWS), Data, AWS Data
Application innovation using AI
Humen.Ai is an AI-driven, synthetic media company. Its popular iOS app, Sway: Magic Dance, is based on AI modeling. The fun-to-use app generates videos of users based on a source video uploaded by the user doing basic motions, like moving around, kicking their legs or waving their arms, for a few seconds. Its PyTorch-backed proprietary machine learning model creates a digital skeleton of the user. Using that skeleton, users can generate a new, photo-realistic stunt double of themselves dancing like Michael Jackson, twirling like a ballet dancer, or making karate moves like a Black Belt in 30 seconds.
The big dance reveals big problems
To operate, Sway: Magic Dance relied on 400 AWS G4 on-demand instances controlled by a complex and expensive-to-maintain system that was built internally. During the 2020 Superbowl, the app was launched in partnership with Doritos©. It hit number two in the Apple App Store and, with help from AWS, Humen.Ai was able to scale up to serve the traffic spike.
The successful launch uncovered scalability and efficiency issues in its backend. The AI app’s concept was built on its ability to quickly train models for each customer and required an immense amount of on demand compute power, as is common in any AI-driven application. “One thing that we were pretty concerned about was the cost of compute in the backend,” said Tinghui Zhou, Co-founder and CEO of Humen.Ai. “Our infrastructure is driven by machine learning, which requires the usage of graphics processing units (GPUs) for model training and inference on the cloud.”
In addition to the infrastructure issues, Humen.Ai faced another common challenge in AI operations. Much like the age-old friction between software developers and engineers, data scientists often face the same friction in moving AI projects from concept through production. AI teams operate much like R&D teams: They are abstracted from engineering once the model is ready to be operationalized. This puts the burden on engineering to figure out architecture design, resource management and model monitoring to run it efficiently, securely and reliably. This friction can slow releases and hinder the pace of innovation. Humen.Ai wanted to remove engineering barriers to reduce the time between building products and enhancements, and releasing them to users while keeping its small, startup team agile.
“Working with a professional team of engineers from Onica on optimizing the backend was overall a very positive experience and really helped us scale our infrastructure.”
Addressing containerization obstacles
The Humen.Ai team was extremely proficient in AI/machine learning operations and in AWS. It had tried a few pathways to optimize infrastructure, like containerization, but those didn’t work out. Amazon SageMaker would have been a great option for model training and inference, as it supports running on Spot Instances, a less expensive route compared to on-demand instances. The tradeoff is needing to wait for AWS to allocate capacity to run spare training, which would hinder the end user experience. Users needed to quickly upload video, have it immediately processed and be able to generate Instagram filters in minutes.
As part of the Jumpstart program, AWS referred Humen.Ai to Onica, a Rackspace Technology® company, for containerization support. The Jumpstart program provides organizations with low-cost infrastructure, credits and training to support growth. Working with Onica, Humen.Ai was able to get to the bottom of its containerization issues. The Onica team helped Humen.Ai containerize its entire AI app to effortlessly deploy on Amazon ECS with Spot Instances.
The Onica team then took it a step further by delivering a custom solution built on AWS. Devising a unique method to schedule Amazon ECS containers, Onica empowered Humen.Ai’s machines to increase the number of AI tasks that could be processed simultaneously. This method, combined with managed services for ECS, enabled it to complete near-real-time training, inference, pre-processing and post-processing using higher-density machines. “We were definitely on the path of switching to a containerized pipeline,” explained Zhou. “Onica put in substantial effort to make that happen for us quickly.”
Exceeding cost-reduction expectations
Combining this extreme software and hardware integration with the use of Spot Instances has resulted in a potential 70% cost reduction — well beyond the expected goal of a 30% reduction. The Humen.Ai team was so enthusiastic about the new infrastructure design that it began building around the project before it was complete. And the Onica team was able to keep up with the speed of the small, agile team, moving from proof of concept to production in just six weeks. Humen.Ai is now leveraging the new lightweight, efficient infrastructure and newfound agility to create more products and reinvest in technology.
Previously, Humen.Ai managed hundreds of instances. By moving to Amazon ECS with Spot Instances, the infrastructure is now easier to manage and operates at lower cost. The AI startup has been able to replace a whole host of infrastructure with managed services. Onica created a lightweight AI infrastructure that allowed Humen.Ai to manage everything using only serverless technologies. This takes a huge burden off of the engineering team and allows the AI team to quickly innovate and efficiently deploy to production.
Costs were reduced by potentially 70% due to the use of Spot Instances combined with extreme software and hardware integration.
Expertise provides the foundation for long-term growth
According to Zhou, “We’ve doubled, maybe even tripled, our throughput of processing demands from our users.” Instances are better utilized, allowing the machines to run more efficiently. This allows the app to handle more concurrent user requests. “I think that definitely had an impact on the user experience,” Zhou continued.
With a very small team, Humen.Ai is now able to build instead of just managing what they’ve already built. It has been able to reduce its technical debt to a level where it’s able to move forward. “We were able to bring down costs significantly and it really improved our back-end efficiency,” noted Zhou as the biggest outcome of his engagement with Onica. The cost reductions were critical in ensuring the early-stage startup’s long-term viability by conserving cash flow.
Reflecting on the experience with Onica, Zhou said, “Working with a professional team of engineers from Onica on optimizing the backend was overall a very positive experience and really helped us scale our infrastructure during the critical, early stages when we don’t have a big, back-end engineering team.”
With the new infrastructure design in place, Zhou plans to expand Humen.Ai’s dance videos to sports, TV, gaming and movies. In addition, the team wants to provide a way for users to share their created content within the app for collaboration. Humen.Ai is also trying to improve the photorealism of its output with 3D-based perception for its skeleton and scene-analysis ML models.
Über Rackspace Technology
Wir von Rackspace Technology sind die Experten für Multicloud-Lösungen. Wir erstellen durch die Kombination unserer Expertise mit den weltweit führenden Technologien lückenlose Lösungen und berücksichtigen dabei Anwendungen, Daten und Security. Wir verfügen über eine nachweisbare Erfolgsbilanz bei der Beratung von Kunden zu ihren individuellen geschäftlichen Herausforderungen. Wir entwerfen, erstellen und verwalten skalierbare Lösungen und optimieren die Rendite mit Blick auf die Zukunft.
Als bahnbrechender globaler Multicloud-Technologieanbieter stellen wir Ihnen die innovativen Fähigkeiten der Cloud zur Verfügung, um Kunden beim Erschließen neuer Umsatzquellen, der Effizienzsteigerung sowie der Schaffung beeindruckender Erlebnisse zu unterstützen. Wir zählen Jahr für Jahr laut Fortune, Forbes und Glassdoor zu den besten Arbeitgebern und fördern erstklassige Talente. So profitieren unsere Kunden von konkurrenzlos guter Expertise. Die Grundlage unseres gesamten Handelns bildet unser unermüdlicher Einsatz für den Erfolg unserer Kunden – unsere Fanatical Experience™. Wir sorgen dafür, dass unsere Kunden schneller und intelligenter arbeiten können und stets einen Schritt voraus bleiben.
Related Case Studies
Let’s Talk Strategy
Tell us a little about your challenges and we’ll contact you.