Founders' Thoughts Featured in War on the Rocks Article
Defense insiders War on the Rocks recently published “Rethinking the Role of a Systems Integrator for Artificial Intelligence,” a piece cowritten by Striveworks founders Jim Rebesco and Tony Manganiello. In it, the two argue that it’s time to pit machine learning models against each other in healthy competition to better serve defense customers.
“The problem is one of competition and the incentives it drives. The market for AI capabilities is uniquely suited to be recast as a continuous competition between models for ‘the right to inference’ individual data points.”
Manage Compute Costs and ML Readiness With New Model-Scaling Configuration
Striveworks customers have even more precision and customization when it comes to scaling model server endpoints.
Chariot users can now configure their model server replicas, allowing them to manage the maximum resources used while maintaining a minimum state of readiness, optimizing both performance and cost efficiency.
Users can also configure—to the minute and on a per-model basis—how long to wait before scaling down resources and the level of traffic at which resources should automatically scale up, ensuring that their AI meets demand in real time without overspending.
Events
Striveworks at AUSA 2024: Overcome Adversarial Adaptation with AI
Increasingly, commanders look to AI models to help make critical, immediate decisions. But environments change, adversaries learn to evade detection, and models fail.
Striveworks’ proven solutions can help your command maintain confidence in those AI models. And if your command is new to AI, we can help your warfighters operationalize AI, quickly and cost-effectively, to meet your pressing needs.
Find us at Booth #538, and schedule a meeting to talk through solutions to your most pressing problems.
Platform Updates
Striveworks Launches Its Segment Anything–Powered Lasso Tool for Easier Segmentation Annotation
Striveworks is making the road to MLOps smoother by incorporating Meta’s Segment Anything Model (SAM) into our platform’s Annotation Studio. Now, users can select objects and parts of objects with a single click, even for overlapping objects.
For anyone who has experienced the time-consuming process of manually annotating complex images by drawing polygons around objects, this upgrade not only simplifies the task but also reduces analyst fatigue, a known challenge to accurate annotation.
Get in Touch
Curious about AI and machine learning projects that scale? Schedule a chat with our team to learn about how we’re making MLOps disappear.
