Hey there š
Iām Jatin, an AGI Enthusiast with a strong desire to conduct research. My primary interests revolve around acquiring an abstract understanding of intelligence with the help of AI. This includes building SOTA algorithms from scratch, creating interpretable networks, and drawing analogies from neuroscience.
Currently pursuing my Master of Science in Computer Science at University of Massachusetts Amherst. I'm working on Mechanistic Interpretability with Prof. David Jenson and a collaborative project with Amazon to understand how LLMs work.
Iām one of the Google DSC Solution Challenge ā21 Global Winner, where I created a system to help Blind and Visually Impaired navigate their surroundings. I have also interned at AdSkate, Inc as a ML Research Intern solving various problems in automated image analysis and Multimodal algorithms.
I have worked in various Machine Learning Domains such as MultiModal AI, MultiDomain AI, Mechanistic Interpretability, Computer Vision, Natural Language Processing, Reinforcement Learning, and Time series Analysis.
You can contact me at: [email protected]
Some of my ongoing projects:
Mechanistic Interpretability for Language Models Biologically Inspired Models and Computation Continual and Lifelong Learning Efficient Deep Reinforcement Learning