Kapil Vaidya

—

I am currently an applied scientist at AWS, where I work on NL2SQL in Amazon Redshift. I was previously a CS PhD in the MIT Data Systems Group, where I work on applying machine learning to database systems. I was advised by Prof. Tim Kraska and I closely collaborated with Prof. Michael Mitzenmacher on learned filters. I have previously interned at Microsoft Research with Anshuman Dutt, Vivek Narasayya and Surajit Chaudhuri from the DMX group on learned query optimisation. I completed my bachelors in Computer Science at IIT Bombay where I worked with Prof. Sudarshan.

[CV] [Google Scholar]

---

Publications

• Vista: Machine Learning based Database Performance Troubleshooting Framework in Amazon RDS.
  Vikramank Singh, Zhao Song, Kapil Vaidya, Balakrishnan Narayanaswamy and Tim Kraska.
  SoCC 2024.

• Why TPC Is Not Enough: An Analysis of the Amazon Redshift Flee.
  Alexander van Renen, Dominik Horn, Pascal Pfeil, Kapil Vaidya, Wenjian Dong, Zhengchun Liu, Gaurav Saxena, Andreas Kipf , Balakrishnan Narayanaswamy and Tim Kraska.
  VLDB 2024.

• Panda: Performance debugging for databases using LLM agents.
  Vikramank Singh, Kapil Vaidya, Vinayshekhar Bannihatti Kumar, Sopan Khosla, Balakrishnan Narayanaswamy, Rashmi Gangadharaiah and Tim Kraska.
  CIDR 2024.

• Can Models Replace Hash Functions?.[news]
  Kapil Vaidya*, Ibrahim Sabek*, Dominik Horn, Andreas Kipf, Michael Mitzenmacher and Tim Kraska.
  VLDB 2023.

• SNARF: A Learning-Enhanced Range Filter.[code]
  Kapil Vaidya, Subarna Chatterjee, Eric Knorr, Stratos Idreos, Michael Mitzenmacher and Tim Kraska.
  VLDB 2022.

• Leveraging Query Logs and Machine Learning for Parametric Query Optimization.
  Kapil Vaidya, Anshuman Dutt, Vivek Narasayya and Surajit Chaudhari.
  VLDB 2022.

• Partitioned Learned Bloom Filter. [Quanta Magazine][code]
  Kapil Vaidya, Eric Knorr, Michael Mitzenmacher and Tim Kraska.
  ICLR 2021.

• The Case for a Learned Sorting Algorithm. [The morning paper] [code] [HackerNews] [Reddit]
  Kapil Vaidya*, Ani Kristo*, Ugur Cetintemel, Sanchit Misra, and Tim Kraska.
  SIGMOD 2020.

• When Are Learned Models Better Than Hash Functions?.
  Kapil Vaidya*, IbrahimSabek*, Dominik Horn, Andreas Kipf, and Tim Kraska.
  AIDB Workshop @ VLDB 2021.

• Defeating duplicates: A re-design of the LearnedSort algorithm. [blog]
  Ani Kristo, Kapil Vaidya, and Tim Kraska.
  AIDB Workshop @ VLDB 2021.

• Ultra-Fast Bit-Level Frequency-Hopping Transmitter for Securing Low-PowerWireless Devices. [news]
  Rabia Tugce Yazicigil, Phillip Nadeau, Daniel Richman, Chiraag Ju-vekar, Kapil Vaidya, and Anantha P. Chandrakasan.
  IEEE Radio Frequency Inte-grated Circuits Symposium (RFIC) 2018.

• Data-Centric Dynamic Partial Order Reduction.
  Marek Chalupa, Krishnendu Chat-terjee, Andreas Pavlogiannis, Nishant Sinha, and Kapil Vaidya
  POPL 2017.

---

Experience

• Research Intern @ MSR Redmond, Summer 2020
• Research Intern @ MIT, Summer 2017
• Research Intern @ IST Austria, Summer 2016

---

Service

• TKDE External Reviewer 2021
• IEEE Transactions on Pattern Analysis and Machine Intelligence External Reviewer 2020

---

Teaching Assistantship

• Database Systems, Spring 2020-21
• Automata Theory, Spring 2017-18
• Digital Logic Design, Spring 2016-17
• CS101, Fall 2016-17

---

Contact

You can contact me by sending a mail to kapilvaidya24@gmail.com.

[Github] [FB]