resumé
Basics
Name | Ashish Rai |
ashish.rai@nyu.edu | |
Url | https://raishish.github.io |
Summary | Interested in Interpretability, Reasoning in Language Models, and AI Safety. |
Work
-
2024.01 - 2025.05 Graduate Assistant - NYU Vital
New York University
Responsible for development and maintenance of the NYU Vital Platform (Private Cloud for assignments).
-
2018.07 - 2023.07 Member of Technical Staff – Performance Engineer
VMware, Bengaluru, India
Optimized resource consumption (CPU, memory) and improved performance of vSAN storage stack.
- Worked on and fixed multiple customer performance bugs in on-prem and cloud environments.
- Led and executed IO path performance characterization for many product feature proposals.
- Contributed to in-house performance frameworks (Python, Java, Angular, Django, SQL).
-
2017.05 - 2017.07 Applied Deep Learning Researcher Intern
Arya.ai, Mumbai, India
Worked on classification of satellite images using Deep Neural Architectures.
- Contributed to a data type agnostic multi-threaded data pre-processing pipeline.
-
2016.12 - 2017.02 Applied IoT Intern
Banyan Nation, Hyderabad, India
Contributed to the development of smart dustbin sensors (based on Intel Edison) to track garbage level.
- Contributed to Analytics and Visualization dashboards (MEAN stack).
Volunteer
-
2020.05 - 2021.08 -
2019.07 - 2020.03
Education
Awards
- 2012
Nominated to attend DST INSPIRE Internship Camp
Top 2 out of ~500.
Skills
Programming Languages & Frameworks | |
Python | |
Pytorch | |
Java | |
SQL | |
LaTeX | |
Angular | |
Django |
Deep Learning Skills | |
Pytorch | |
HuggingFace | |
DDP | |
Slurm |
Other Skills | |
Virtual Machines | |
Cloud (AWS) | |
Distributed Computing | |
Teaching |
Languages
English | |
Fluent |
Hindi | |
Native Speaker |
Projects
- 2024.11 - 2024.11
- 2024.10 - 2024.12
A vision-based approach to ARC-AGI
Trying vision and multimodal models to advance the ARC-AGI challenge
- 2024.04 - 2024.05
Human vs Machine Attention in Visual Question Answering
Predicting answers from images and comparing human vs. model attention.
- 2023.11 - 2023.12
Self-supervised stochastic conditional video prediction
Given a set of context frames, predict the next set of frames of videos.Done as part of the final project for CSCI-GA 2572 - Deep Learning.