Sitemap
A list of all the posts and pages found on the site. For you robots out there, there is an XML version available for digesting as well.
Pages
Posts
Why is Alignment Hard?
Published:
What we mean by alignment, what people have tried so far, and why we are still stuck.
Why AI Safety?
Published:
Convincing myself (and maybe you) why I am pursuing AI safety — technical safety, with a hint of governance and philosophy.
portfolio
Portfolio item number 1
Short description of portfolio item number 1
Portfolio item number 2
Short description of portfolio item number 2
publications
Systematic Biomarker Discovery for Glioblastoma Subtyping Using Machine Learning
Published in 2025 IEEE 6th India Council International Subsections Conference (INDISCON), 2025
A systematic machine learning pipeline for discovering biomarkers that distinguish glioblastoma subtypes using transcriptomics data, feature selection, and gene co-expression network analysis.
Recommended citation: Sutradhar, A., Banerjee, A., Samanta, A., Hossain, S. F., Dey, A., & Biswas, S. (2025). "Systematic Biomarker Discovery for Glioblastoma Subtyping Using Machine Learning." IEEE INDISCON 2025. pp. 1–6.
Download Paper
Learning Filter-Aware Distance Metrics for Nearest Neighbor Search with Multiple Filters
Published in arXiv preprint arXiv:2511.04073, 2025
We propose a method to learn filter-aware distance metrics that improve approximate nearest neighbor retrieval efficiency and accuracy under multiple conjunctive attribute filters.
Recommended citation: Sutradhar, A., Gupta, S., Krishnaswamy, R., Xu, H., Rastogi, A., & Srinivasa, G. (2025). "Learning Filter-Aware Distance Metrics for Nearest Neighbor Search with Multiple Filters." arXiv preprint arXiv:2511.04073.
Download Paper
WinSyn: An Automated Pipeline for Realistic Enterprise Question-Answering Evaluation
Published in arXiv preprint arXiv:2609.12171, 2026
An automated pipeline that generates synthetic enterprise email corpora with grounded questions and reference answers, used to evaluate agentic question-answering systems on realistic, long-running workplace scenarios.
Recommended citation: Varhade, A., Sutradhar, A., Krishnaswamy, R., & Goyal, N. (2026). "WinSyn: An Automated Pipeline for Realistic Enterprise Question-Answering Evaluation." arXiv preprint arXiv:2609.12171.
Download Paper
Side Effects of Character Training: Quantifying Cross-Constitution Drift in LLMs
Published in ICML 2026 Workshop on Pluralistic Alignment, 2026
We quantify how character training a language model toward one constitution induces unintended behavioral drift when the model is evaluated against other constitutions.
Recommended citation: Kumar, B., Sutradhar, A., Panigrahi, S., Chang, J., & Levine, L. (2026). "Side Effects of Character Training: Quantifying Cross-Constitution Drift in LLMs." ICML 2026 Workshop on Pluralistic Alignment.
Download Paper