Kinjal Bhavsar

Bootstrapped and delivered SceneCore, the inaugural developer SDK for Android XR as part of Jetpack XR — the foundation for building immersive 3D applications on Android devices.

• Designed and launched robust input handling mechanisms and an Entity Component System (ECS) in SceneCore, enabling expressive spatial app development for Android XR.
• Serve as API reviewer and Kotlin expert for the SceneCore library — reviewing 250+ CLs, enforcing API council standards, Kotlin idioms, and library design consistency.
• Designed and implemented the entity lifecycle management system — DisposedException, isDisposed, automatic cleanup via XrReferenceCleaner, eliminating resource-leak bugs in spatial apps.
• Owned ActivitySpace spatial continuity: gravity-aligned unscaled activity space, FSM-gated origin updates, yaw preservation, and worldspace scale overrides.
• Expanded the public SceneCore API surface: unrestricted InteractableComponent, InputEvent, GltfModelEntity; shipped Kotlin Flow + lifecycle-aware listener extensions.
• Led Java → Kotlin migration of the SceneCore runtime — converting 20+ core classes including SpatialSceneRuntime, MovableComponentImpl, SpatialEnvironmentImpl, and more.
• Modernized the SceneCore test app to MVVM + Jetpack Compose, replacing all XML layouts and adding ViewModel infrastructure.

Worked on ARCore, Google's augmented reality platform powering AR experiences across billions of Android devices worldwide.

• Led targeted power optimizations in ARCore's camera pipeline — critically enabling the launch and viability of Project Guideline, Google's AR navigation app for people with low vision.
• Spearheaded a cross-functional team to design and launch Electronic Image Stabilization (EIS) for ARCore — gyroscope-driven stabilization correcting rolling shutter and camera shake in real time across Android SDK and NDK. docs →
• Re-architected and re-implemented the ARCore camera data pipeline, improving device compatibility and enabling new camera features including a Kotlin-based camera stack.
• Designed and deployed a comprehensive analytics framework to gather and analyze ARCore camera performance data across an install base of over 2 billion devices.

• Developed a framework enabling non-VR Android applications to launch and operate within the GearVR environment, extending the GearVR content ecosystem.
• Designed and prototyped novel AR experiences for internal evaluation and future product consideration.
• Implemented CameraX extensions for Samsung devices, providing third-party developers access to advanced camera features (HDR, Night Mode) on Samsung Galaxy hardware.

• Hired and mentored a platform software team; led successful porting of the Qualcomm camera stack to a Xenomai real-time kernel and associated RT libraries.
• Directed development of custom drivers for large 1-inch, 16-bit HD image sensors integrated with Qualcomm SoCs.
• Drove real-time implementation of 3A algorithms (Auto-Exposure, Auto-White Balance, Auto-Focus) and significant camera data path optimization — delivering substantially reduced photon-to-display latency.

Contributed to the world's first array camera for consumer mobile devices, enabling post-capture refocus and depth sensing.

• Defined the architecture of the camera software stack from drivers to APIs, ensuring system cohesion and efficiency.
• Developed a critical array sensor driver for capturing high-quality raw image data across MIPI CSI-2 interfaces.
• Expanded the Android camera framework to export depth data, unlocking new developer capabilities; optimized ISP pipeline performance.

• Delivered significant optimizations to the Android display stack (SurfaceFlinger, Hardware Composer) for Snapdragon chipsets.
• Ported and optimized the Android bootloader (LK) for enhanced boot performance and stability on new Snapdragon platforms.
• Profiled Adreno graphics driver performance and provided OEM partner support, enabling successful device launches.

Board bring-up and BSP development for a Qualcomm MSM7x50-based platform — Linux kernel initialization, display (framebuffer) and input (touchscreen) drivers for AOSP.

Developed a framework and device drivers for real-time audio/video transcoding on TI DaVinci and DSP platforms, enabling efficient real-time media processing.

Advanced computer systems: microprocessor architecture, OS design, embedded systems, RTOS, networking and internet protocols, parallel computing, digital ASIC design, and memory systems.

Bachelor of Engineering in Electronics & Communication Engineering from one of Gujarat's premier technical universities, formerly known as D.D. Institute of Technology.

Mentored middle school students in mathematics, helping build confidence and foundational skills in STEM subjects.