Signals
Vision is the interface between AI and the physical world. These posts cover computer vision as it actually ships — production perception, AR, try-on, and the hard lessons that separate demos from deployed systems.
Every imaging system — an eye, a telescope, a phone camera, a computer vision pipeline — runs into the same wall: physics sets a hard limit on resolution. A tour of where that limit comes from.
Build a working 478-landmark face tracker in a single HTML file — MediaPipe from a CDN, getUserMedia, and a canvas overlay. No npm, no bundler, working code included.
The gap between a computer vision demo and a production system is where most projects die. Lessons from two decades of shipping perception systems to real users in the real world.
Face, body, and object tracking; segmentation; depth and gaze estimation — tuned for real-time performance on the hardware you actually ship on.
Photorealistic real-time compositing for eyewear, apparel, beauty, and jewelry — a category Ashwin helped define across global retail.
Embedding-based visual search, style and attribute recognition, and systems that connect what a camera sees to what a catalog contains.
Marker tracking, alignment, and measurement for medical and industrial domains — where sub-millimeter accuracy is the requirement.
Real-time 3D perception and virtual overlay deployed across CHANEL's worldwide retail presence.
Next-generation virtual try-on combining spatial tracking with real-time garment simulation.
Award-winning real-time AR experience at Walt Disney World and Hong Kong Disneyland. Millions of guests interacting with intelligent visual systems.
Systems designed and led by Ashwin Rajendraprasad in prior roles — shown as experience, not client work of CloudSignal AI.
Live hand and face tracking, AR try-on, and visual search — running entirely in your browser. Check back, or write to us for an early look.
Tell us what your system needs to see — we'll tell you honestly whether it's feasible, and what it takes.