Sanjay Ghemawat

Ghemawat was born in 1966 in West Lafayette, Indiana, and grew up in Kota, an industrial city in northern India. He did not touch a computer until he entered Cornell University at the age of seventeen.⁠[1][1]

In 1987, he received a Bachelor of Science degree (S.B.) from Cornell University, then went on to graduate study at the Massachusetts Institute of Technology (MIT), earning a master’s degree (S.M.) in 1990.⁠[2][2]

In September 1995, he submitted his doctoral thesis, “The Modified Object Buffer: A Storage Management Technique for Object-Oriented Databases,” to MIT’s Department of Electrical Engineering and Computer Science, studying storage management for object-oriented databases under thesis supervisors Barbara Liskov and Frans Kaashoek. Liskov studied the management of complex code bases and held that the best code is like a good piece of writing, with a carefully realized structure in which every word does work.⁠[2][2][2][2][1][1][1]

Before joining Google, Ghemawat worked at the Systems Research Center of Digital Equipment Corporation (DEC), while Jeff Dean was at DEC’s Western Research Lab two blocks away; the two had already begun working together, often coding at the same computer. Dean joined Google first, and Ghemawat followed in December 1999. He has said that he might never have interviewed at Google if Dean had not been there.⁠[3][1][1][1][3]

In March 2000, the core systems that Google used to crawl the web and build its index had stopped working properly since the previous October, leaving search results five months out of date, while the company was negotiating to power search for Yahoo. Ghemawat and Dean joined a makeshift “war room” to investigate; on the fifth day, they converted the corrupted index file to its rawest binary form and found that a digit that should have been a 0 was a 1. The problem lay in the machines’ corrupted memory chips, and they then wrote code to compensate for the faulty machines, allowing the new index to be completed.⁠[1][1][1][1][1][1][1][1][1]

The two then led Google’s effort to build a fault-tolerant computing system out of large numbers of inexpensive machines: they wrote code so that a single hard drive could fail without bringing down the entire system, added checkpoints to the crawling process so that it could be restarted midstream, and, by developing new encoding and compression schemes, effectively doubled the system’s capacity. In 2001, over four days, they proved that Google’s index could be stored in memory rather than on hard drives, reshaping the company’s cost structure. Over the next three or four years, the two, together with an ever-changing group of engineers, rewrote Google’s core systems for crawling the web, building the index, and returning search results several times.⁠[1][1][1][1][1][3]

In 2003, over four months, Ghemawat and Dean gave Google what was arguably its biggest single upgrade with MapReduce. The idea came to them the third time they rewrote Google’s crawler and indexer: they realized that each time they had been solving the same problem, coordinating work across a vast number of distributed, individually unreliable computers. Programmers need only write two functions, “map” and “reduce,” while the system handles partitioning the data, scheduling tasks, and dealing with failures. The following year, the two rewrote the crawling and indexing system using MapReduce and presented a paper at OSDI 2004 listing only the two of them as authors. Developers Doug Cutting and Mike Cafarella built an open-source clone, Hadoop, from scratch based on it, which was later widely adopted.⁠[1][1][1][4][4][1][3][3][4][1][1]

In October 2003, Ghemawat, as first author with Howard Gobioff and Shun-Tak Leung, presented “The Google File System” at the SOSP conference, describing the distributed file system (GFS) Google designed for large-scale data-intensive applications: it runs on inexpensive commodity hardware, provides fault tolerance, and had been widely deployed as the storage platform for data generation and processing at Google.⁠[5][5][5][5][5]

In 2005, Ghemawat and Dean helped develop Bigtable, a system for treating large numbers of computers as a single database, after which a cafeteria on Google’s campus was later named. In 2006, he presented “Bigtable: A Distributed Storage System for Structured Data” at OSDI with Fay Chang, Dean, and others, describing a structured data storage system that scales to petabytes of data across thousands of commodity servers and is used to store data by projects including web indexing, Google Earth, and Google Finance.⁠[1][6][6][6][6]

In 2012, Ghemawat was among the authors of “Spanner: Google’s Globally-Distributed Database,” presented at OSDI, describing Google’s scalable, multi-version, globally distributed, and synchronously replicated database. The same year, the ACM Special Interest Group on Operating Systems (SIGOPS) gave the Mark Weiser Award to him and Dean.⁠[7][7][7][8]

The preliminary TensorFlow white paper released on November 9, 2015, listed Ghemawat as one of its authors. TensorFlow is a machine learning framework whose development was led by Dean, and it was released to the public the same year.⁠[9][9][1][1]

In 2016, Ghemawat was inducted into the American Academy of Arts and Sciences.⁠[1]

As of 2018, Ghemawat and Dean were Google’s first and only two Level 11 engineers, with the title of Google Senior Fellow. At the time, he was an “individual contributor” who managed no one, working on software to make it easier for engineers to combine and control the dozens of programs behind a search request, and he met weekly with the “Area Tech Leads” group that made technical decisions for the entire company; Dean had time to code with him only once a week, and during their Monday coding sessions the two started an AI project, which Dean said aimed to train a “giant” machine-learning model able to do thousands or millions of different tasks.⁠[1][1][1][1][1][1][1]

On August 5, 2026, Ghemawat, Jeff Dean, Quoc Le, and Oriol Vinyals announced that they were leaving Google to co-found Discovery Loop, a public benefit corporation that aims to use AI to automate complete experimental loops, initiating and iterating thousands of experiments simultaneously to accelerate engineering and scientific discovery, starting with the automation of machine learning research. The initial funding round was co-led by Radical VenturesRadical VenturesRadical Ventures is an artificial intelligence-focused venture capital firm founded by Jordan Jacobs and Tomi Poutanen shortly after the 2017 launch of the Vector Institute in Toronto. It invests from incubation and seed through growth and pre-IPO stages, and its portfolio includes companies such as Cohere, Waabi, and World Labs. In 2026, it co-led the investment in Discovery Loop, founded by Jeff Dean and others, and completed the first close of its late-stage Radical Breakouts Fund.Open the full entry and Khosla VenturesKhosla VenturesKhosla Ventures (KV) is an American venture capital firm founded in 2004 by Vinod Khosla, the founding CEO of Sun Microsystems, and headquartered in Menlo Park, California. It made heavy early bets on clean technology, became OpenAI’s first venture capital investor in 2019, and its portfolio also includes companies such as DoorDash, Block (formerly Square), Impossible Foods, and Rocket Lab.Open the full entry, with participation from Kleiner Perkins, Lightspeed, and Doerr Capital, and Google also invested and agreed to provide compute power for the first year. Google CEO Sundar Pichai said that over 27 years, Jeff and Sanjay had helped drive a series of major technology transitions, from early search infrastructure to neural networks.⁠[10][11][10][10][10][10][12][10][11][11]

Ghemawat and Dean make a habit of coding together at the same computer, with Ghemawat typically doing the typing; standard engineering practice is to have code reviewed by another programmer, but the two skip this step, entering a perfunctory “lgtm” (looks good to me) in their log. Dean considers Ghemawat more level-headed and able to correct their course, while Ghemawat says Dean keeps them moving forward. Their colleague Wilson Hsieh said Dean is great at coming up with wild new ideas and prototyping, while Ghemawat was the one who built things to last.⁠[3][1][1][3][3][3][1]

Ghemawat and Dean are the authors of LevelDB, Google’s open-source key-value storage library, which provides an ordered mapping from string keys to string values.⁠[13][13]

Discovery Loop: Official website: https://www.discoveryloop.com/ (opens in a new window).⁠[12]