Hyderabad: Blue Cloud Softech Solutions Limited (BSE: 539607) has announced that it has secured a major international deal from Stratos Forge Inc, marking a significant milestone for the company’s Blue Cloud Softech Data Annotation Project portfolio.
The agreement follows the successful completion of a pilot phase valued at approximately ₹18.00 crore, during which Blue Cloud delivered a high annotation accuracy of 96.68% across agreed KPIs.
Full-Scale Rollout of Blue Cloud Softech Data Annotation Project
Following the strong pilot performance, Stratos Forge Inc has awarded Blue Cloud Softech a full-scale project with an estimated commercial value of ₹110.08 crore.
Also Read: Deepwatch Opens Bangalore Center to Accelerate AI-Powered Cybersecurity Innovation
The project will be executed through Blue Cloud’s in-house delivery ecosystem, supported by Centres of Excellence (CoEs) established in collaboration with top-tier universities. This structure enables the company to manage high-volume, high-precision AI data workflows.
Blue Cloud’s Advanced Annotation Techniques for the Project
The Blue Cloud Softech Data Annotation Project will adopt advanced, automation-assisted annotation methodologies designed for enterprise-grade AI systems. These include:
AI-Assisted and Automated Annotation
- Active Learning: AI models identify uncertain samples, reducing manual labeling by up to 90%.
- Model-Assisted Pre-Labeling: Pre-trained models generate initial annotations, delivering 5x–10x workflow acceleration.
- Weak Supervision: Programmatic labeling using frameworks like Snorkel to generate large-scale probabilistic labels.
- Synthetic Data Generation: Using GANs and Diffusion Models to produce privacy-safe, auto-labeled synthetic datasets.
- Advanced Quality Control Frameworks
- Inter-Annotator Agreement (IAA): Metrics such as Cohen’s Kappa ensure statistical consistency.
- Gold Set Validation: Embedded ground-truth samples continuously monitor annotator accuracy.
- Semantic Consistency Checks: Automated logic-based validation to prevent inconsistent labels.
Optimized Workflow Models
- Human-in-the-Loop (HITL): Continuous prediction–correction–retraining cycles for maximum accuracy.
- Micro-Tasking: Decomposition of complex tasks into focused steps to reduce cognitive load.
Also Read: Ness Digital Engineering to Build AI-Enabled Product Engineering Ecosystem for Vendavo
Domain-Specific Expertise
Blue Cloud’s domain specialists and academic partners will apply tailored annotation solutions across:
- Autonomous mobility
- Robotics and industrial vision
- Natural language processing
- Behavioral intelligence
- 3D LiDAR and point-cloud analysis
Janaki Yarlagadda, Chairman of Blue Cloud Softech Solutions Ltd, said: “We are delighted to expand our partnership with Stratos Forge Inc after our highly successful pilot engagement. Their confidence in Blue Cloud highlights the strength of our annotation automation frameworks, our CoE talent pipeline, and our ability to deliver world-class AI training data at scale.”
A spokesperson for Stratos Forge Inc added: “Blue Cloud demonstrated exceptional technical precision and execution quality during the pilot phase. This extended engagement marks the next phase of our collaboration to build highly accurate datasets powering our advanced AI products and enterprise systems.”
