About Me
Hello there! I am a second-year Master’s student at Stanford University’s Institute of Computational and Mathematical Engineering (ICME). I am broadly interested in Artificial Intelligence, specifically in building systems that can reliably and robustly learn to reason.
In 2022, I spent a wonderful summer interning as an Applied Scientist at Amazon where I worked on permutation-invariant Large Language Models to refine the knowledge graph at the backend of the Amazon Catalogue System. In my first year at Stanford, I worked on a wide variety of projects, ranging from Deep Generative Models, to domain-agnostic Self-Supervised Learning and Natural Language Processing.
Before joining Stanford, I worked as a Quantitative Researcher at a high-frequency trading firm in India where I led the automation and backtesting of trading strategies using Statistical Machine Learning. In 2020, I graduated with a B.Tech after spending 4 wonderful years as an undergraduate at IIT Guwahati, majoring in Mathematics and Computer Science.
If you find my work interesting, or want to discuss about some cool projects that you are working on, feel free to reach out at aamdekar at stanford dot edu
- Natural Language Processing
- Computer Vision
- Graph Neural Networks
M.S, Computational and Mathematical Engineering, 2023
Stanford University
B.Tech, Mathematics and Computer Science, 2020
Indian Institute of Technology, Guwahati
Experience
- Owned an end-to-end service that leveraged state-of-the-art permutation invariant Transformers and Graph Neural Networks to refine the knowledge graph that supports the Amazon Catalogue System.
- Designed a model that was 2.5x more computationally efficient than the production model and achieved a significant gain on out-of-distribution tasks.
- Collaborated with Prof. Stefano Ermon on diffusion models using score matching for representation learning in JAX.
- Worked on domain-agnostic Self-Supervised Learning algorithms in Pytorch.
