Murshed Al Amin
Data Scientist @ AIDocBuilder / InteliWeave Inc
I’m a Data Scientist at Inteliweave, where I build machine learning systems for large-scale document intelligence, spanning dataset design, model development, deployment, monitoring and performance optimization. My work sits at the intersection of engineering and research, contributing to reliable AI pipelines used in real-world, production environments.
As a researcher, I’m broadly interested in machine learning, with an ongoing focus on mechanistic interpretability. I previously conducted a mechanistic interpretability study on ODEformer as part of AI Safety Camp 2025, where I investigated internal representations in time-series transformers.
I completed my B.S. in Computer Science from Independent University, Bangladesh, where I also conducted undergraduate research in graph neural networks.
In my free time, I enjoy watching films, traveling, and discussing life’s existential questions with friends.
latest posts
| Nov 27, 2025 | Understanding & Implementing a Tiny GPT-2 (With Limited GPU) |
|---|---|
| May 28, 2025 | AlexNet Paper Implementation |
| Jul 15, 2024 | Mechanistic Interpretability via Learning Differential Equations |
selected publications
-
Data Quality Assessment of Substation Data in Bangladesh: Insights from Handwritten Data DigitizationIn 2023 10th IEEE International Conference on Power Systems (ICPS), 2023 -
Assessment of Rehabilitation Exercises from Depth Sensor DataIn 2021 24th International Conference on Computer and Information Technology (ICCIT), 2021