Anubhav Goel

Hey I am Anubhav, a Masters in Computer Science student at the University of Texas at Austin, where I am working as a GRA under Prof. Haris Vikalo. I love studying Deep Learning techniques, understand the math behind them and build systems to deploy them for real world problems. I am currently working at Futurewei Technologies and building a 3D rendering system for mixed reality applications.

I am researching the capabilities of transformer-based deep learning models and their applications in fields beyond natural language, specifically if they can understand long-term correlations in the human genome. More specifically, I am using them for early detection of cancer in human beings.

I completed my Bachelors in Electrical Engineering at the Indian Institute of Technology, India where I worked with Prof. Vivek Borkar, with whom I explored the theoretical methods in the field of Reinforcement Learning and Stochastic Optimization as well as Prof. Suyash Awate, with whom I explored generative modeling methods and their application for the anomaly detection problem in medical image processing.

In 2021, I visited Technical University of Denmark on an exchange program where I had the wonderful opportunity to collaborate with people from all over the world. In 2020, I interned at McKinsey & Co. where I worked on building an end-to-end data analytics tool for generating insights on revenues and investment strategies for banks. Here is my CV for more information on my past projects.

Updates

October 2023: Here's a collaborative drawing tool that I made using sockets.
May 2023: I will be joining Futurewei Technologies as an ARVR Algorithm Research Intern this summer!
January 2023: I have been offered a GRA position!
August 2022: I will be a Teaching Assistant under Prof. Joydeep Ghosh for the Advanced Machine Learning course at UT Austin.
June 2022: I received J N Tata scholarship to pursue graduate studies!
June 2022: I was awarded Excellence in Academics at IIT Bombay!
February 2022: I was awarded the Excellence in Teaching Assistantship award at IIT Bombay!
January 2022: I started working as a Teaching Assistant for Machine Learning under Prof. Amit Sethi.
December 2021: I received an offer to join for a full time position at Morgan Stanley!
January 2022: I started working as a Teaching Assistant for Image Processing under Prof. Amit Sethi.
January 2021: I started my exchange semester at the Technical University of Denmark.
September 2020: I started working as a Teaching Assistant for Differential Equations II under Prof. Amiya Kumar Pani.
August 2020: Completed my Data Analyst internship at McKinsey & Co.!

Projects

Controlled Generation using Diffusion-LM and FUDGE: We performed a comparative study between two popular controlled-generation models (and several other baselines) to evaluate the deployability of large language models in real-life.
[report] [code]
GNN: A Survey on Architectures and Optimization: We wrote a term paper on different GNN architectures and various optimzations to speed them up.
[report]
Optimizing cp -r: We used io_uring released in Linux 5.1 to wrote a faster implementation of cp -r.
[report]
Disagreement in GNN Explanation Methods: We studied different methods for explaning the output of GNNs and how they disagree for a given problem as well as ways to quantify that disagreement.
[report] [code]
Deep Learning Strategies for Reconstructing Undersampled RS fMRI: I compiled a survey on the various deep learning techniques used for reconstruction of undersampled fMRI images.
[report]
3D Scene Reconstruction Using Multiple 2D Images: We extracted SIFT features from various 2D images and used a FLANN matcher to reconstruct the correspondin 3D scene.
[presentation]
U-Net for Segmentation of Neuronal Structures in EM Stacks: We explored the U-Net architecture for image segmentation.
[poster]
Rendering and Animation using OpenGL: We created a short animated movie using OpenGL. We used cool concepts such as hierarchical modeling, Phong lighting model, texture mapping and skyboxes!
[report] [movie]
StyleGAN for audio: We adapted the popular StyleGAN algorithm for the audio domain.
[report]
Learning Bridges
[code]
Toonification
[presentation]
Implementation of RISC in VHDL
15 Puzzle
[code]