April 2026 Summaries

Creatium has revolutionized online learning by creating immersive, AI-powered educational experiences that leverage pedagogical science and real-time AI systems to enhance productivity for instructional designers. However, as the platform evolved, infrastructure limitations became a significant challenge, causing provisioning delays and reliability issues. To address these hurdles, Creatium evaluated several infrastructure providers and ultimately chose Cerebrium for its ability to support both training and inference workloads efficiently. The implementation of Cerebrium's serverless infrastructure allowed for dynamic GPU provisioning, significantly reducing cold-start times to about 10 seconds and improving overall performance and reliability. This transition eliminated the need for prior workarounds, reduced operational complexity, and allowed engineers to focus on enhancing the learning product, all while achieving 99.999% uptime and cost efficiency.

Lelapa AI is dedicated to advancing language technologies for African languages, particularly in transcription and translation, but faced challenges with infrastructure, specifically cold-start times of up to 30 minutes, which hampered their ability to provide real-time services. Despite initially using Hugging Face for deployment, the inefficiencies led them to explore alternatives, eventually choosing Cerebrium for its effective developer experience, responsive customer support, and pay-for-what-you-use model. The switch to Cerebrium dramatically reduced cold-start times and error rates, allowing Lelapa to focus on their core mission without being bogged down by infrastructure issues. This transformation not only improved their operational efficiency but also strengthened their partnership with Cerebrium, enabling Lelapa to better serve African communities and expand their language coverage.

Tavus, a company focused on developing empathetic AI-human interactions, faced significant challenges in scaling their infrastructure, particularly with GPU deployments for their Custom Video Infrastructure workloads. The team experienced issues with cold starts, inflexible auto-scaling, and traffic surges, which affected user latency and deployment efficiency. After evaluating various platforms, Tavus chose Cerebrium due to its responsiveness and quick onboarding process, which significantly reduced development cycle times and enhanced product agility. This choice allowed Tavus to deploy changes rapidly, maintain reliability during high-demand periods, and scale their operations seamlessly, giving them a competitive edge. The support from Cerebrium, especially during unexpected issues, along with favorable infrastructure spending, has been crucial in allowing Tavus to focus on their core mission of integrating empathy into technology.

Distil Labs, a developer platform for building task-specific small language models with high accuracy, faced challenges in maintaining cost-effective and scalable infrastructure for model deployment and inference. To address these challenges, they partnered with Cerebrium, which provided a comprehensive platform solution that enabled dynamic scaling, optimized cold starts, and competitive pricing. This partnership allowed Distil Labs to focus on improving their models and customer value, while Cerebrium handled the infrastructure needs, including autoscaling and global deployment capabilities. As a result, Distil Labs achieved significant improvements in inference cost and model accuracy, while maintaining consistent latency and reliability, allowing them to handle high-traffic periods effectively. The collaboration with Cerebrium also fostered a highly responsive and integrated working relationship, further enhancing Distil Labs' operational efficiency.

bitHuman is transforming human-device interaction by developing intelligent digital avatars capable of real-time emotional engagement, particularly in omni-channel sales and customer interactions. The company, led by CEO Steve Gu, stands out due to its on-device performance, generative flexibility, and offline-first capability, allowing models to run entirely on CPUs without cloud dependency, making deployments more economical and efficient. Initially, bitHuman faced challenges with Azure's infrastructure, such as long cold-start times and unreliable autoscaling, which hindered their growth. A pivotal shift occurred when they adopted Cerebrium, reducing launch times significantly and enhancing deployment efficiency, which allowed for 24/7 service availability with substantial cost savings and improved reliability. This partnership not only simplified deployment processes but also empowered the team to focus on creating digital characters that offer natural, instantaneous interactions, positioning bitHuman for scalable future growth.