CV
Ananya Sutradhar
Summary
Anthropic Fellow working on AI safety and alignment research. Previously Pre-doctoral Research Fellow at Microsoft Research India (Team DiskANN), focused on making retrieval better and evaluating Deep Research systems. SPAR Fellow at Cornell with Prof. Lionel Levine.
Education
- Computer Science and Technology2025Indian Institute of Engineering Science and Technology (IIEST) ShibpurGPA: 9.5
Work Experience
- Anthropic Fellow2026-05-01 -Working on AI safety and alignment research.
- Pre-doctoral Research Fellow, Team DiskANN2025-01-01 - 2026-05-01Worked on making retrieval better and evaluating Deep Research systems.
- Software Development Engineer Intern2024-05-01 - 2024-07-31Implemented domain-driven design principles to optimize the notification triggering process.
Skills
Machine Learning
- Supervised learning
- Metric learning
- Feature selection
- Retrieval evaluation
Algorithms & Systems
- Approximate nearest neighbor search
- Vector search
- Data structures
- Algorithm design
Programming
- C
- C++
- Python
- Kotlin
- Rust
- JavaScript
Tools & Frameworks
- PyTorch
- TensorFlow
- scikit-learn
- AWS
- SQL
- Git
Publications
- Side Effects of Character Training: Quantifying Cross-Constitution Drift in LLMs2026ICML 2026 Workshop on Pluralistic AlignmentQuantifies how character training a language model toward one constitution induces unintended behavioral drift when evaluated against other constitutions. Work done as a research mentor for SPAR.
- Systematic Biomarker Discovery for Glioblastoma Subtyping Using Machine Learning2025IEEE INDISCON 2025A systematic pipeline for biomarker discovery in glioblastoma subtyping using transcriptomics data and machine learning, incorporating feature selection and gene co-expression network analysis.
- Learning Filter-Aware Distance Metrics for Nearest Neighbor Search with Multiple Filters2025arXiv preprintWe study the problem of nearest neighbor search under multiple conjunctive filters and propose a method to learn filter-aware distance metrics that improve retrieval efficiency and accuracy.
Interests
- Approximate Nearest Neighbor Search
- Metric Learning
- Information Retrieval
- Evaluation of Deep Research Systems
- AI Safety