doi.bio/jerry_tworek


Jerry Tworek

Early Life and Education

Jerry Tworek, also known as Jaroslaw Tworek, is a research scientist and lead at OpenAI. He attended the University of Warsaw, where he obtained a Bachelor of Science degree in Applied Mathematics between 2008 and 2012. He continued his studies at the same university from 2012 to 2013, earning a Master of Science in Applied Mathematics.

Career

Tworek began his career as a software engineering intern at Google, where he worked on gathering and analyzing latency data and conducted original research on collaborative filtering algorithms for Google's social products. He then took on roles at various financial institutions, including JP Morgan Chase and Decura, where he gained experience in quantitative analysis, algorithmic trading, and investment strategy development.

In 2019, Tworek joined OpenAI as a research scientist in neural program synthesis. He focused on neural computer program synthesis research, reinforcement learning, and reasoning with large language models. He also contributed significantly to ChatGPT and GPT-4. Currently, he leads research on teaching LLMs to use tools and solve complex problems in STEM fields.

Scientific Contributions

Tworek has made notable contributions to the field of AI and machine learning. Some of his scientific works include:

Talks

Tworek has also given talks on language models and their capabilities, such as his presentation on "Language Models Are Few-shot learners," where he discussed the research behind GPT-3 and its engineering challenges.

Jerry Tworek

Early Life and Education

Jerry Tworek, or Jarosław Tworek, is a research scientist and lead at OpenAI, specialising in deep reinforcement learning.

Tworek attended the University of Warsaw, graduating with a Bachelor of Science in Applied Mathematics between 2008-2012, and later returning to obtain a Master of Science in Applied Mathematics from 2012-2013.

Career

Tworek started their career as a software engineering intern at Google, where they worked on latency data gathering and analysis, and original research on collaborative filtering algorithms for Google social products. They then moved to JP Morgan Chase as an equity derivatives structuring intern, developing a platform for algorithmic tradable indices.

In 2013, Tworek became a quantitative analyst at Decura and worked as a realtime graphics engineer at Platige Image. They then moved on to become a senior quantitative researcher at Independent View, where they focused on developing new quantitative investment strategies and risk management.

In 2019, Tworek joined OpenAI as a research scientist in neural program synthesis, working on neural computer program synthesis research, reinforcement learning, and reasoning with large language models. They have made major contributions to ChatGPT and GPT-4, and their team built the Plugins and Code Interpreter.

Tworek has also presented their work at NeurIPS 2019 DeepRL Workshop, and given talks on GPT-3 and its capabilities, as well as the research and engineering challenges involved in training large language models.

Scientific Contributions

Tworek's scientific contributions include:

Youtube Videos

Youtube Title: Reinforcement Learning agent playing Atari Enduro

Youtube Link: link

Youtube Channel Name: Jerry Tworek

Youtube Channel Link: https://www.youtube.com/@jerrytworek9044

Reinforcement Learning agent playing Atari Enduro

Youtube Title: Reinforcement Learning agent playing Atari Breakout

Youtube Link: link

Youtube Channel Name: Jerry Tworek

Youtube Channel Link: https://www.youtube.com/@jerrytworek9044

Reinforcement Learning agent playing Atari Breakout

Youtube Title: Warsaw ai Episode IX 22.10.2020

Youtube Link: link

Youtube Channel Name: Warsaw AI

Youtube Channel Link: https://www.youtube.com/@warsawai

Warsaw ai Episode IX 22.10.2020

Youtube Title: Warsaw ai - Episode XIII, 14.10.2021

Youtube Link: link

Youtube Channel Name: Warsaw AI

Youtube Channel Link: https://www.youtube.com/@warsawai

Warsaw ai - Episode XIII, 14.10.2021

Youtube Title: Codex: Large Language Models Trained on Code

Youtube Link: link

Youtube Channel Name: Data Science Gems

Youtube Channel Link: https://www.youtube.com/@dlByManish

Codex: Large Language Models Trained on Code

Youtube Title: Reinforcement Learning agent playing Atari Beamrider

Youtube Link: link

Youtube Channel Name: Jerry Tworek

Youtube Channel Link: https://www.youtube.com/@jerrytworek9044

Reinforcement Learning agent playing Atari Beamrider

Youtube Title: Midjourney v6, Altman 'Age Reversal' and Gemini 2 - Christmas Edition

Youtube Link: link

Youtube Channel Name: AI Explained

Youtube Channel Link: https://www.youtube.com/@aiexplained-official

Midjourney v6, Altman 'Age Reversal' and Gemini 2 - Christmas Edition

Youtube Title: Reinforcement Learning agent playing Atari Qubert

Youtube Link: link

Youtube Channel Name: Jerry Tworek

Youtube Channel Link: https://www.youtube.com/@jerrytworek9044

Reinforcement Learning agent playing Atari Qubert

Youtube Title: Yi Large: Surprisingly Great at Logic and Reasoning! (Fully Tested)

Youtube Link: link

Youtube Channel Name: Matthew Berman

Youtube Channel Link: https://www.youtube.com/@matthew_berman

Yi Large: Surprisingly Great at Logic and Reasoning! (Fully Tested)

Youtube Title: AI Learning to Play Tom & Jerry: Reinforcement Q-Learning

Youtube Link: link

Youtube Channel Name: Abdurrahman TEKİN

Youtube Channel Link: https://www.youtube.com/c/AbdurrahmanTEK%C4%B0N

AI Learning to Play Tom & Jerry: Reinforcement Q-Learning

Youtube Title: More For Les: A Living Tribute - Patrick Tworek aka Bloody Jerry

Youtube Link: link

Youtube Channel Name: Les Hernandez

Youtube Channel Link: https://www.youtube.com/@leshernandez5358

More For Les: A Living Tribute - Patrick Tworek aka Bloody Jerry

Youtube Title: I Made a Language Learning AI (you can too)

Youtube Link: link

Youtube Channel Name: Jerry Registre

Youtube Channel Link: https://www.youtube.com/@JerryRegistre

I Made a Language Learning AI (you can too)

Youtube Title: BERT-QE: Contextualized Query Expansion for Document Re-ranking (Research Paper Walkthrough)

Youtube Link: link

Youtube Channel Name: TechViz - The Data Science Guy

Youtube Channel Link: https://www.youtube.com/@TechVizTheDataScienceGuy

BERT-QE: Contextualized Query Expansion for Document Re-ranking (Research Paper Walkthrough)

Youtube Title: Long term credit assignment with temporal reward transp… | Cathy Yeh | OpenAI Scholars Demo Day 2020

Youtube Link: link

Youtube Channel Name: OpenAI

Youtube Channel Link: https://www.youtube.com/@OpenAI

Long term credit assignment with temporal reward transp… | Cathy Yeh | OpenAI Scholars Demo Day 2020

Youtube Title: Stunning AI shows how it would kill 90%. w Elon Musk.

Youtube Link: link

Youtube Channel Name: Digital Engine

Youtube Channel Link: https://www.youtube.com/@DigitalEngine

Stunning AI shows how it would kill 90%. w Elon Musk.

Youtube Title: Solving Rubik's Cube with a Robotic Hand Explained!

Youtube Link: link

Youtube Channel Name: Connor Shorten

Youtube Channel Link: https://www.youtube.com/@connorshorten6311

Solving Rubik's Cube with a Robotic Hand Explained!

Youtube Title: All about OpenAI chatGPT how to register / Login,how to ask questions and save it .

Youtube Link: link

Youtube Channel Name: Online Learning Junction

Youtube Channel Link: https://www.youtube.com/@onlinelearningjunction6764

All about OpenAI chatGPT how to register / Login,how to ask questions and save it .

Youtube Title: LIVE: Tesla's unveils a masterpiece: The Tesla that will change the car industry forever - Tesla CEO

Youtube Link: link

Youtube Channel Name: Flict-G

Youtube Channel Link: https://www.youtube.com/@news.motors24

LIVE: Tesla's unveils a masterpiece: The Tesla that will change the car industry forever - Tesla CEO