About Me

I am a PhD Student at the Data Systems Lab of MIT. My research focuses on building practical, performant, and cost-effective cloud systems and on reliably programming complex systems with the help of Large Language Models.

Systems

Mallet: SQL Dialect Translation with LLM Rule Generation

Mallet is a system that uses LLMs to automatically read SQL system documentations, analyze the semantics of their dialects, and build a pre-defined set of rules to translate SQL queries between different dialects. It aims to be scalable, accurate, fast, and human-verifiable.

BRAD: Data Mesh Optimization

BRAD aims to unify and jointly optimize (performance-wise and cost-wise) multiple specialized database engines in a workload-dependent manner.

VISC: Transcending the Serverless - Serverful Dichotomy

Virtualized Serverless Computing (VISC) unifies cloud programming APIs behind a single interface, facilitating the use of multiple underlying serverless or serverful implementations. For a given workload, VISC automatically balances these underlying implementations to reach a performance target in a cost-optimal manner.

Publications

Mallet: SQL Dialect Translation with LLM Rule Generation
Amadou Latyr Ngom, Tim Kraska
Venue: SIGMOD 2024 -- AIDM Workshop
Blueprinting the Cloud: Unifying and Automatically Optimizing Cloud Data Infrastructures with BRAD
Geoffrey X. Yu, Ziniu Wu, Ferdi Kossmann, Tianyu Li, Markos Markakis, Amadou Ngom, Samuel Madden, and Tim Kraska.
Venue: VLDB 2024
Check Out the Big Brain on BRAD: Simplifying Cloud Data Processing with Learned Automated Data Meshes
Tim Kraska*, Tianyu Li*, Samuel Madden*, Markos Markakis*, Amadou Ngom*, Ziniu Wu*, Geoffrey Yu*
Venue: VLDB 2023 -- Vision Paper
*: Equal Contribution
Filter Representation in Vectorized Query Execution
Amadou Ngom, Prashanth Menon, Matthew Butrovich, Lin Ma, Wan Shen Lim, Todd C. Mowry, Andrew Pavlo
Venue: SIGMOD 2021 -- DAMON Workshop.
Permutable Compiled Queries: Dynamically Adapting Compiled Queries Without Recompiling
Prashanth Menon, Amadou Ngom, Matthew Butrovich, Lin Ma, Wan Shen Lim, Todd C. Mowry, Andrew Pavlo
Venue: VLDB 2021
Mainlining Databases: Supporting Fast Transactional Workloads on Universal Columnar Data File Formats
Tianyu Li, Matthew Butrovich, Amadou Ngom, Wan Shen Lim, Wes McKinney, Andrew Pavlo
Venue: VLDB 2021
Everything is a Transaction: Unifying Logical Concurrency Control and Physical Data Structure Maintenance in Database Management Systems
Ling Zhang, Matthew Butrovich, Tianyu Li, Andrew Pavlo, Yash Nannapaneni, John Rollinson, Huanchen Zhang, Ambarish Balakumar, Daniel Biales, Ziqi Dong, Emmanuel J Eppinger, Jordi E Gonzalez, Wan Shen Lim, Jianqiao Liu, Lin Ma, Prashanth Menon, Soumil Mukherjee, Tanuj Nayak, Amadou Ngom, Dong Niu, Deepayan Patra, Poojita Raj, Stephanie Wang, Wuwen Wang, Yao Yu
Venue: VLDB 2021

Teaching

Software Systems for Data Science -- MIT
Spring 2022 TA
Database Systems -- Carnegie Mellon University
Fall 2019 TA
Distributed Systems -- Carnegie Mellon University
Spring 2019 Head TA, Fall 2018 TA, Spring 2018 TA