Английский Технологии
27.08.2026 Читать источник
Основатели Reducto назвали стартап заклинанием из «Гарри Поттера» для превращения документов в данные
Выпускники MIT Адит Абрахам и Раунак Човдури создали компанию Reducto, названную в честь заклинания из «Гарри Поттера», которая преобразует сложные документы в структурированные данные для искусственного интеллекта. Стартап, оцененный в 600 миллионов долларов, использует технологию, аналогичную магии из книг, чтобы извлекать информацию из рукописного текста и мелких деталей, недоступных обычным сканерам.
Перевод ИИ
Текст статьи доступен на языке оригинала.
Нажмите кнопку «Перевести статью» выше, чтобы сгенерировать перевод с помощью ИИ.
Оригинальный контент
Harry Potter movies and novels have been an integral part of childhood, sending children into a fantasy world. But it seems that some MIT alumni have taken this childhood fantasy to another level.
Adit Abraham and Raunak Chowdhuri, two MIT graduates of Indian origin, named their AI company Reducto after the "Reductor Curse," a spell in the books that blasts solid objects into fragments.
It's a fitting bit of trivia, because what their startup actually does is quite less dramatic but useful in the AI era.
Meet Adit Abraham and Raunak Chowdhuri, the Indian origin MIT alumni who named their startup after a Harry Potter spell
Abraham and Chowdhuri's story goes back to their time as computer science students at MIT. According to Adit Abraham, in an interview on the First Round Review podcast, his earliest memory of Raunak dates back to when he was a junior at MIT and had just signed up for his first graduate-level machine learning course.
He recalled feeling a fair amount of imposter syndrome walking in, and it was on that very first day that the professor introduced the class to Raunak. That early connection eventually turned into a working partnership. The two stayed in touch through their time at MIT and later decided to build something together, applying to the startup accelerator Y Combinator with an idea that had nothing to do with documents at all.
Their original pitch to Y Combinator wasn't Reducto; it was a tool for giving large language models long-term memory.
What is their educational background?
Abraham leads Reducto's vision and strategy, driving its document-to-data platform and key growth initiatives. He holds a B.S. in Computer Science from MIT, specialising in machine learning. He previously worked as a Product Manager at Google, co-founded Sidewalk Labs, and was an ML Engineer at BlinkAI.
Chowdhuri looks after Reducto's technical strategy, overseeing architecture and machine learning development. He has a B.Eng. from MIT focused on AI and robotics, and previously served as CTO at Oloren AI, an ML researcher with MIT Driverless, and an ML consultant at MIT Lincoln Laboratory.
What is Reducto and how did it get its name?
According to Forbes, Reducto has grown into a company valued at $600 million, backed by more than $100 million in funding, and used by major names like Scale AI, Vanta, and Airtable.
However, the Reducto name wasn't a calculated branding decision. Speaking to Fortune, Abraham explained, "We had this notion of wanting to build magical tools for developers. We had a list of names that were magic-adjacent. Reducto was one of them, and we picked it. Weirdly, it works with what we do, but it wasn't at all intentional."
The name comes from the "Reductor Curse" in the Harry Potter series, a spell that blasts objects into smaller pieces or dust. Reducto's actual technology does something conceptually similar to documents. The software takes large, complex files and breaks them down into structured, accurate, LLM-ready data, including handwriting, checkboxes, and other tiny markings a typical scanner would miss.
The pair's work also earned them a spot on Forbes' 2026 30 Under 30 list for artificial intelligence.
Also Read: Meet Lakshmanan Meyyappan: The 16-year-old boy who set a world record as the youngest Chartered Accountant in the world
Аналитика ИИ и парсинга
Технические детали обнаружения, сопоставления и связывания статьи
Поисковый запрос
«Reducto стартап документы в данные ИИ»
Запрос, сгенерированный ИИ для поиска статьи в веб-источниках
Связанные результаты поиска
Aug 11, 2024 · В заключение С помощью нового API от Reducto вы легко можете преобразовывать сложные документы и таблицы в структурированные данные, совместимые со схемой, без необходимости ручной настройки.
Reducto AI AI-API для обработки документов, который преобразует сложные неструктурированные документы в высокоточные структурированные данные, готовые для LLM.