AI & Machine Learning Product Engineering Case Studies

Cardinal Peak/FPT engineered an AI-powered IDP solution for a global chemical leader to automate data extraction from 300+ complex Safety Data Sheets daily. By combining advanced computer vision with NLP, we achieved 90+% accuracy and a 90% reduction in manual effort.

Cardinal Peak/FPT engineered a large-scale AI video analytics platform for a Japanese telecommunications conglomerate to secure university campuses nationwide. By developing custom Person Re-identification (Re-ID) models and a hybrid edge-cloud vision architecture, we enabled seamless tracking of individual trajectories across disconnected camera networks at over 20 distributed locations.

Cardinal Peak/FPT engineered a custom AI inventory management solution for a rapidly growing pharmacy chain to scale from 4 to 1,000+ locations. By deploying predictive demand forecasting and an AI semantic search engine, we reduced stock-outs to under 5% and drove a 2.5x increase in online revenue.

Cardinal Peak/FPT engineered a custom computer vision and Edge AI solution for NXP Semiconductors to automate microscopic wafer defect classification. By optimizing 11 deep learning models for ARM-based edge devices, we achieved 95% accuracy and an 80% reduction in labor costs while maintaining under 1-second inspection latency.

Cardinal Peak/FPT engineered a high-performance edge-based biometric access system for a global ICT leader to replace slow manual verification. By optimizing facial recognition for low-power edge compute and integrating with physical security hardware, the solution achieved 99%+ accuracy and a frictionless “walk-through” experience for 16,000 employees globally.

A top-tier automotive supplier partnered with us to engineer a deep learning-based vision system for high-precision, reflective components. By overcoming lighting glares and surface reflections that traditional vision systems miss, our custom solution achieved a zero-escape rate for critical defects and a 30% reduction in inspection time.

Cardinal Peak / FPT engineered a cloud-native deep learning pipeline to automate high-content screening of B-cells. By replacing manual microscopy analysis with automated instance segmentation and classification, the solution achieved a 10x increase in throughput and a 30% improvement in identifying usable cells for antibody discovery.

We engineered a custom AI-powered acoustic test fixture for a global appliance manufacturer to automate subjective manual inspections. By utilizing deep learning and advanced DSP to isolate motor defects amidst 90dB factory noise, the system achieved 95% accuracy and a 100% increase in inspection throughput within the first month.

By optimizing AI workflow automation with AI, video and AWS-powered solutions, Cardinal Peak helped WorkDone cut processing costs by 99%, enhance procurement efficiency and scale seamlessly. Read the full case study to see how artificial intelligence-driven automation transforms business processes.

Faced with customers filling terabyte drives and struggling to collaborate, our client needed to centralize storage, accelerate processing and enable seamless teamwork. Discover how Cardinal Peak built a multitenant cloud solution on AWS that streamlines digital image processing and facilitates easy collaboration through cloud access.

In this case study, Cardinal Peak’s video product design and development expertise unlocked remote accessibility and significantly more efficient intel surveillance, helping one of New Jersey’s largest private investigation firms realize an astounding 80% profit boost.

This case study reveals the secret to outsmarting mischievous feline antics with our innovative cat detector powered by Amazon Kinesis Video Streams, AWS IoT Greengrass and machine learning.