Ivan Bercovich

Partner

2020 - Current

We focus on post-revenue companies close to the $500k-1M milestone. Among my partners, I focus primarily on tech/AI companies, and general operating support for our founders.

Board

2021

CEO

2022 - 2023

I joined full time to help HeyTutor's transition from startup to growth business. Alongside the executive team, we adopted stricter procedures and executed meaningful organizational changes. In a year, we grew the business by 5x.

Board

2022

Board

2022

Board

2021

Board

2021

Co-Founder (2021), Investor (2022+)

2021

I was the a co-founder of Unwrap, a startup we incubated at the Allen Institute for AI. We focused on making sense of customer feedback at scale. I was involved in building the first prototype and raising early capital, but ultimately decided to pursue investing full time. Unwrap is one of our investments at ScOp VC.

Entrepreneur In Residence

2021 - 2021

GM for Semantic Q&A | Alexa

2017 - 2020

Amazon acquired our company, Graphiq, to improve Alexa's open domain question understanding and answering (Q&A) capabilities. At the time, 2017, Alexa was behind competitors in the Q&A space, but assisted by the Graphiq acquisition, Alexa caught up to be on par or be the best-in-class assistant on various knowledge categories. I led the team of 130 software developers and knowledge engineers, who jointly built the technology and knowledge graph supporting Alexa's semantic Q&A capabilities. By the time I ended my tenure in 2020, the Graphiq technology was answering most Q&A questions across all languages (English, German, Japanese, Spanish, French, Portuguese, Hindi) and locales in which Alexa was present. The technology our team built takes a transcribed input from automatic speech recognition (ASR) and handles all the natural language processing, knowledge retrieval, and natural language generation, and then hands off the final text to a text to speech system (TTS). In parallel, we built a number of internal tools and workflows that enabled our knowledge engineers to define complex knowledge representation via a proprietary ontology and ingested billions of facts to build one of the most comprehensive knowledge graphs (KG) in the world. Additionally, we made some of this technology available to customers through Alexa's Structured Knowledge Skills offering. Amazon 2018 shareholder letter: "Last year, we improved Alexa’s ability to understand requests and answer questions by more than 20%, while adding billions of facts to make Alexa more knowledgeable than ever."

Advisory Board

2016 - 2019