Geometrically Guaranteeing the Accuracy of Cache Hits in Top-K ANN Search for Vector-Indexed Databases
2026papercodepresentation
~/anupta >
Anupta Argo
Fourth-year CS student at Princeton. Interested in distributed and database systems, the intersection of systems and ML, and large-scale infrastructure.
About
Hi, I'm Anupta, an incoming MS CS student at Princeton. My research focuses on database systems and large-scale infrastructure. I'm drawn to open-ended problems where rigorous systems design meets real-world constraints.
Lately I've been especially interested in how ML and systems can inform each other: both using ML to optimize systems internals, and building systems that better support ML workloads. I'm excited to keep exploring this through graduate research and collaborations.
Research
Current and past research.
Evaluating Mixing Ratios in Simultaneous Live Audio Translation Systems
2025papercodepresentation
Ablating Semantic Underspecification
2025Projects
Selected and in-progress projects.
Mini-Milvus
2026
Implementing a lightweight vector database system in C++ inspired by Milvus, including core ANN indexing and retrieval primitives, query execution, and storage/runtime components for high-dimensional vector search.
Raft Consensus
2025
From-scratch implementation of the Raft consensus protocol in Go with leader election and log replication. Built a fault-tolerant distributed key-value store on top.
PrincetonNotes
2025
Decentralized note-sharing platform for Princeton courses where students and course staff can create, upload, and moderate content.
TigerRooms
2024
A COS333 project that simplifies Princeton's housing draw by parsing room availability PDFs into a clean, searchable interface.
TigerTrain
2024
Study platform for Princeton's COS226 with LeetCode-style practice, answer checking, and AI-assisted help.