Scaled AI compute 10x to power frontier cancer research
Reduced sample processing time by 80-90%, freeing scientists for discovery
Cut ML experiment cycles from 3-4 weeks to just one week
Accelerates new oncology therapeutics from lab to patient
Enabled AI-driven research services not offered by traditional labs
Hematogenix built an AI research engine on Google Cloud to pioneer precision oncology and bring discoveries to patients faster.
Hematogenix built an AI research engine on Google Cloud to pioneer precision oncology and bring discoveries to patients faster.
A patient-first approach to precision medicine
For a patient waiting for a cancer diagnosis or treatment plan, the faster they get results, the faster they can begin life-saving therapies. Hematogenix, a life science company that’s driven by a science and patient-first mission, was founded to accelerate this timeline, specializing in delivering precision medicine solutions and proprietary diagnostic clinical oncology assays for patients around the world.
With a global research and development team supported by more than 100 PhDs and MDs, the company is constantly adapting its clinical assays and developing proprietary AI models to quickly bring impactful and more affordable cancer therapeutics to patients. However, as the business focused heavily on new AI initiatives, its fixed, on-premises compute infrastructure became a significant bottleneck.
Through more than 18 years of cancer diagnostic work, the company has built and curated a proprietary real-world multimodal dataset of millions of samples and over a trillion data points. It needed a highly scalable solution that could support massive training runs without compromising strict data governance. Hematogenix chose Google Cloud for its established reputation, comprehensive HIPAA, GDPR, and ISO compliance, and strategic alignment with life sciences.
We've collected and structured one of the world’s largest proprietary multimodal diagnostic sets of oncology data. Google Cloud provides a single environment that delivers both scalable compute and the built-in compliance that we need to manage that kind of scale.
Bishr Almasri
Director of Innovation, Hematogenix
"We've collected and structured one of the world’s largest proprietary multimodal diagnostic sets of oncology data,” says Bishr Almasri, Director of Innovation at Hematogenix. “Google Cloud provides a unified environment that delivers both scalable compute and the built-in compliance that we need to manage that kind of scale.”
Scaling compute and streamlining AI experimentation
By transitioning to Google Cloud, Hematogenix gained access to automated infrastructure scaling, which replaces the limitations of fixed-cost on-premises machines. Now, the company provisions resources on demand and seamlessly spins them down when jobs are completed. “Our pipelines scale based on our workload,” Almasri says. “That helps us stay lean and efficient, but also gives us that freedom to scale up as high as we may need.”
Google Cloud's extensive data storage and compute resources allowed us to build capacity and infrastructure that scales with us as we go from experiments to training to production-ready pipelines.
Bishr Almasri
Director of Innovation, Hematogenix
Using Google Cloud's data and AI tools, including BigQuery and Gemini, Hematogenix has radically streamlined its research pipelines. Machine learning engineers use Workbench to rapidly curate datasets in BigQuery and run small-scale experiments, such as predicting genetic mutations. Once validated, they easily convert the workbench and spin up additional resources to move these experiments into full production pipelines.
“Google Cloud's extensive data storage and compute resources allowed us to build capacity and infrastructure that scales with us as we go from experiments to training to production-ready pipelines,” Almasri says.
This transition has transformed the pace of research and development, turning processes that previously took three to four weeks using on-premise resources into one-week turnarounds. With Google Kubernetes Engine and the Model Garden, Hematogenix has also significantly enhanced engineering productivity, allowing for easier resource provisioning and parallel project execution.
Accelerating the frontier of cancer research
Partnering with Google Cloud has enabled Hematogenix to scale its compute capacity by nearly 10x, upgrading from 16 on-premise GPUs to hundreds on demand. Meanwhile, custom in-house models trained on Google Cloud have driven an 80% to 90% reduction in turnaround time for critical sample processing workflows.
These operational efficiencies have a profound, real-world impact. Faster workflow processing means that oncology patients receive their test results quicker, enabling them to begin life-saving therapies sooner.
Looking ahead, Hematogenix plans to use this scalable infrastructure to pioneer AI-driven clinical research, focusing on accelerating the analysis of tissue biopsies and whole slide images to conserve precious samples and reduce overall costs. With a modernized, AI-ready foundation, Hematogenix is successfully overcoming the boundaries of big data to bring treatments to patients who need them most.
“We're building some of the most ambitious oncology AI models in the world—models that will eventually impact millions of people globally once they're available,” Almasri says. “Google Cloud has given us the secure, scalable, AI-ready infrastructure to accelerate cancer diagnostics and pioneer the future of precision medicine.”
Google Cloud has given us the secure, scalable, AI-ready infrastructure to accelerate cancer diagnostics and pioneer the future of precision medicine.
Bishr Almasri
Director of Innovation, Hematogenix
Hematogenix specializes in delivering precision medicine solutions and proprietary diagnostic oncology assays for patients worldwide, taking a science-first approach to accelerate cancer research.
Industry: Healthcare and Life Sciences
Location: United States
Products: Google Cloud, BigQuery, Compute Engine, Gemini, Google Kubernetes Engine, Model Garden on Gemini platform
