Hey, I'm Anjali!

I'm a creative technologist and machine learning engineer at Meta Reality Labs, where I work on Codec Avatars — 3D body tracking, evaluation/curation, and the data pipelines that power them. I hold an M.S. and B.S. in EECS from UC Berkeley, with a concentration in computer vision.

I'm drawn to work that sits at the intersection of the physical and the digital. My interests span 3D reconstruction, computational color, and computer graphics, and I'm especially excited about the ways sensing and rendering technology can enable new kinds of experiences. I have previously worked at Disney Imagineering R&D, UCSB Media Arts and Tech Lab, and the UCSF Big Data in Radiology Lab.

Currently: developing interactive exhibits with the Exploratorium New Media Group, and learning projection mapping.

For more info about my experience, click here! LinkedIn

Feel free to check out a few of my projects below :)

Graphics

Sand-Water Simulation

Implemented multi-species (sand & water) particle interactions using the 2D Material Point Method (Taichi Graphics, Python)

Rain Simulation

Simulated raindrops sliding down a glass window - modeled raindrop shape, motion, and shading.

Cloth Simulation

Implemented a real-time simulation of cloth using a mass and spring based system (OpenGL, C++)

Rasterizer

Implemented a simple rasterizer including features like drawing triangles, supersampling, hierarchical transforms, and texture mapping with antialiasing. (OpenGL, C++)

Mesh Editing

Built Bezier curves and surfaces using de Casteljau algorithm, manipulated triangle meshes represented by half-edge data structure, and implemented loop subdivision.

Path Tracer

Implemented physically-based renderer using a pathtracing algorithm. Specifically, implemented ray-scene intersection, acceleration structures, and physically based lighting and materials

Extension of Path Tracer

Implemented mirror and glass BSDFs as well as a depth of field simulation

Computer Vision

Facial Keypoint Detection with Neural Nets

Created an automatic facial keypoint detector (68 keypoints) from scratch using Pytorch.

Neural Style Transfer

Reproduced the algorithm and results as described in the paper "A Neural Algorithm of Artistic Style paper" by Leon A. Gatys, Alexander S. Ecker, and Matthias Bethge.

Object Pose Estimation

Implemented pose estimation of a known image. Adapted this pose estimation algorithm to compute a sparse point cloud representation of a scene. (OpenCV, Numpy, Python)

Implementation of Neural Network from Scratch

Computational Photography

Processing Prokudin-Gorskii Glass Plate Images

Use digitized versions of Prokudin-Gorskii’s RGB glass plate negatives to faithfully reconstruct a RGB colored image.

Image processing: Hybrid Images & Multi-Resolution Blending

Used high and low frequencies of an image to blend, sharpen, and combine a set of images in different ways.

Face Morphing

Used user-defined facial keypoints to combine, morph, and warp faces in interesting ways!

Auto-Stitching Photo Mosaics

Stitched together images taken from different angles to create one contiguous parnorama shot. Used principles of homogenous matrices, image warping, and image combination / blending

Image Quilting

Implemented a contiguous, "quilted" image using a single texture patch, as well as texture transferring.

Light Field Camera

Implemented effects like depth refocusing and aperture adjustment using collections of large image sets taken over a plane that is orthogonally aligned to the optical axis.