Anish Thite

Education

Georgia Institute of Technology, Atlanta GA

M.S. in Computer Science (Intelligence and Devices threads) - Expected May 2022

B.S. in Computer Science - May 2021

Overall GPA (B.S.): 3.78/4.00 | Major GPA: 4.00/4.00

Skills

Languages: Python, Java, Rust, C, Scala, Bash, JS

Tools/OS: Git, AWS, GCP, Firebase, Docker, Linux, Windows, SQL, Slurm, CUDA

Frameworks: Flask, Django, React, PyTorch, TensorFlow, Numpy, Pandas

Experience

Amazon AWS Migrations Hub Team

Software Design Intern, Boston, MA (May 2021 - July 2021)

• Working on a data generation service for testing AWS Migration Hub recommendations

Georgia Tech Research Institute

Graduate Research Assistant, Atlanta, GA (January 2019 - Present)

• Creating General Adversarial Networks for waveform generation and simulation
• Using genetic programming to evolve machine learning models with the GTRI Evolutionary Auto-ML Framework
• Expanding the library to perform neural network architecture search (using Python, MySQL, Scikit-learn, Keras, Tensorflow)
• Out-performed Google Auto-ML on the Wikidetox toxic comment classification dataset
• Student Workshop Paper "Concurrent Neural Tree and Data Preprocessing AutoML for Image Classification" accepted at GECCO 2021

The Agency (AI/ML Club)

External Relations Officer (President), Atlanta, GA (January 2020 - May 2021)

• Sys admin for club server, lecture on NLP topics, coordinate guest lecturers, and advise project groups

EleutherAI

Researcher/Software Engineer, Atlanta, GA (July 2020 - Present)

• Contributed to replicate and open-source GPT-3 weights, wrote open-source codebase for language model evaluation
• Working on HTML + PDF text extraction and filtering to make The Pile, one of the world's largest test datasets
• Efficiently scraped, filtered, and parsed entire Hacker News for language model training
• https://arxiv.org/abs/2101.00027
• https://github.com/EleutherAI/lm_evaluation_harness

NLP and Computational Social Science Lab

Student Researcher, Atlanta, GA (November 2019 - Present)

• Working under Dr. Diyi Yang on keyword-based neural humor generation and classification using transformers
• Further study includes analyzing effects of humor on mental health and the COVID pandemic

High-Performance Architecture Lab

Student Researcher, Atlanta, GA (August 2019 - December 2019)

• Researched in performing deep learning tasks on edge
• Incorporated speech recognition and conversation models (cosine-similarity response matching) on Raspberry Pis