doi.bio/robert_verkuil

Robert Verkuil

Robert Verkuil is a scientific researcher and a PhD student working under Dr. [Name of PI] at [Name of University/Institute].

Research Interests and Publications:

Verkuil's research focuses on computational biology and machine learning, specifically in the application of these fields to protein structure and function prediction and understanding.

He has co-authored several papers with other researchers, including:

• Roshan Rao, Jason Liu, and others, on the "MSA Transformer," published in ICML 2021.
• Joshua Meier, Roshan Rao, and others, on "Language models enable zero-shot prediction of the effects of mutations on protein function," published in NeurIPS 2021. This paper had one code implementation.
• Chloe Hsu, Jason Liu, and others, on "Learning inverse folding from millions of predicted structures," published in ICML 2022.

Other Projects:

Verkuil has also worked on several other projects, as evidenced by his GitHub repositories. These include:

• "Chrome-Trace-Perf-Tool": A tool related to Chrome browser performance tracing.
• "drake-pytorch": Combining PyTorch (Python and C++ bindings) with Drake for neural network systems.
• "Button-Machine": A 2nd front button machine, using node.
• "TensorFlow-Examples": TensorFlow tutorial and examples for beginners, forked from aymericdamien/TensorFlow-Examples.
• "TACO-TensorFlow-Integration": Details not provided.

Education and Academic Background:

[Provide details of Verkuil's academic background, including where he obtained his Bachelor's and Master's degrees, and the year he started his PhD.]

Awards and Recognition:

[List any awards or notable recognition received by Verkuil for his research contributions.]

Contact:

Verkuil can be contacted via his GitHub profile: robert-verkuil [link to profile], or through his Twitter handle: @robert_verkuil.

References:

1. Robert Verkuil. (n.d.). https://github.com/robert-verkuil
2. Robert Verkuil. (n.d.). https://paperswithcode.com/search?q=author:Robert+Verkuil
3. Robert Verkuil. (n.d.). https://dblp.org/pid/296/8930.html

Robert Verkuil

Overview

Robert Verkuil is a researcher and student at the Massachusetts Institute of Technology (MIT). He has a Bachelor of Science (B.S.) in Electrical Engineering and Computer Science from MIT, and has taken courses in Advanced Natural Language Processing, Artificial Intelligence, Circuits and Electronics, and Machine Learning.

Work Experience

Verkuil has held several internships and applied machine learning roles, gaining experience in software engineering, electrical engineering, and machine learning model development.

Internships

Applied Machine Learning (AML) - Translation

• Developed a rapid prototyping pipeline for testing deep neural translation models.
• Determined that models retained accuracy with significantly fewer weight bits.
• Co-developed parallelization methods to exploit memory bandwidth savings.
• Implemented changes that reduced average translation time by 35%.

Hudson River Trading

• Worked on metadata server upgrades.
• Co-developed a tool for generating parsers for structured text files.

Five Rings Capital, LLC

• Designed and tested a high-speed, mixed-signal, low-noise prototype PCB.
• Managed external contractors and communicated with vendors and in-house experts.

Software Engineering Intern

• Upgraded the metadata server's throttling system.

Platform Architecture Intern

• Created a calibration rig and GUI for in-factory programming of 3D-printer parts.

Embedded Firmware Intern

• Reduced power consumption for idling products by 30%.
• Identified wireless chip inefficiencies, resulting in additional power savings.
• Developed a predictive battery life model for Pebble watches, accurate within ±5%.

Electrical Engineering Intern

• Utilised 3D printing to create product enclosures, brackets, and gears for prototypes.
• Designed and assembled surface mount circuit boards.
• Developed a computer-controlled power supply for stress testing and optimising battery usage.

Publications

Verkuil has co-authored several publications, including:

• "Learning inverse folding from millions of predicted structures" (ICML 2022)
• "MSA Transformer" (ICML 2021)
• "Language models enable zero-shot prediction of the effects of mutations on protein function" (NeurIPS 2021)

GitHub

Verkuil has a GitHub account with several public repositories:

• Chrome-Trace-Perf-Tool: A tool for Chrome tracing and performance analysis.
• drake-pytorch: Combining PyTorch with drake for neural network systems.
• Button-Machine: A 2nd front button machine project.
• TensorFlow-Examples: TensorFlow tutorial and examples for beginners.
• TACO-TensorFlow-Integration: Repository for TACO-TensorFlow integration.

ResearchGate

On ResearchGate, Verkuil is credited with contributions to the following works, among others:

• "Evolutionary-scale prediction of atomic-level protein structure with a language model"
• "Language models generalize beyond natural proteins"
• "Learning inverse folding from millions of predicted structures"

- "Language models enable zero-shot prediction of the effects of mutations on protein function"

Robert Verkuil

Overview

Robert Verkuil is a researcher and student at the Massachusetts Institute of Technology (MIT). He holds a Bachelor of Science (B.S.) in Electrical Engineering and Computer Science from MIT. Verkuil has had multiple internships in software engineering and machine learning, and has contributed to several research papers in the fields of machine learning and protein structure prediction.

Education

• Massachusetts Institute of Technology – B.S. in Electrical Engineering and Computer Science
• University of Virginia School of Law – J.D.
• New School for Social Research – M.A.
• New York University School of Law – LL.M. and J.S.D.

Work Experience

• Intern, Applied Machine Learning (AML) – Translation (Jun/2017-Aug/2017) – Menlo Park, California
• Algorithm and Core Developer Winter Intern, Hudson River Trading (Jan/2017-Feb/2017) – Greater New York City Area
• Software Engineering Intern (Jun/2016-Aug/2016)
• Platform Architecture Intern, Five Rings Capital, LLC (Jun/2015-Aug/2015)
• Calibration Hardware Intern (Jan/2015)
• Embedded Firmware Intern (Jun/2014-Aug/2014)
• Electrical Engineering Intern (Jun/2013-Oct/2013)

Research Contributions

Verkuil has co-authored several research papers in the fields of machine learning and protein structure prediction. His contributions include:

• Learning inverse folding from millions of predicted structures: This paper proposes a method for predicting protein sequences from backbone atom coordinates using machine learning. By augmenting training data with predicted structures, the authors achieve improved performance in sequence recovery.
• Language models enable zero-shot prediction of the effects of mutations on protein function: The paper demonstrates that protein language models can capture the functional effects of sequence variations without any supervision from experimental data, performing at a state-of-the-art level.
• MSA Transformer: The paper introduces a transformer model that utilizes sequence alignment information to improve protein modeling.
• Evolutionary-scale prediction of atomic-level protein structure with a language model: This work demonstrates the use of large language models to directly infer full atomic-level protein structure from primary sequences, enabling large-scale structural characterization of metagenomic proteins.

References

1. Verkuil, R., & others (2022). Learning inverse folding from millions of predicted structures. ICML, 8946-8970.
2. Rao, R., Liu, J., Verkuil, R., Meier, J., Canny, J. F., Abbeel, P., … & Rives, A. (2021). MSA Transformer. ICML, 8844-8856.
3. Meier, J., Rao, R., Verkuil, R., Liu, J., Sercu, T., & Rives, A. (2021). Language models enable zero-shot prediction of the effects of mutations on protein function. NeurIPS, 29287-29303.
4. Verkuil, R., & others (n.d.). Evolutionary-scale prediction of atomic-level protein structure with a language model. ResearchGate. Retrieved April 2, 2024, from https://www.researchgate.net/scientific-contributions/Robert-Verkuil-2189872855

Youtube Videos

Youtube Title: Video1

Youtube Link: link

Youtube Channel Name: RobertVerkuil

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

Video1

Youtube Title: Intra-Body Communicator

Youtube Link: link

Youtube Channel Name: RobertVerkuil

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

Intra-Body Communicator

Youtube Title: Robo See Robo Do

Youtube Link: link

Youtube Channel Name: RobertVerkuil

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

Robo See Robo Do

Youtube Title: Video2

Youtube Link: link

Youtube Channel Name: RobertVerkuil

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

Video2

Youtube Title: Local Professor Gives Testimony at Trump Trial

Youtube Link: link

Youtube Channel Name: WLFITV

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

Local Professor Gives Testimony at Trump Trial

Youtube Title: Zandvoort Onboard 9-10-2020 1:499

Youtube Link: link

Youtube Channel Name: Thomas Verkuijl

Youtube Channel Link: https://www.youtube.com/@BMWE36Thomas

Zandvoort Onboard 9-10-2020 1:499

Youtube Title: Interoperability: A Foundation for the Future of Education

Youtube Link: link

Youtube Channel Name: 1EdTech

Youtube Channel Link: https://www.youtube.com/@1EdTech

Interoperability: A Foundation for the Future of Education

Youtube Title: Robert Young - Metric Embeddings and the Geometry of the Heisenberg Group

Youtube Link: link

Youtube Channel Name: Plectics

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

Robert Young - Metric Embeddings and the Geometry of the Heisenberg Group

Youtube Title: Linus Torvalds Guided Tour of His Home Office

Youtube Link: link

Youtube Channel Name: The Linux Foundation

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

Linus Torvalds Guided Tour of His Home Office

Youtube Title: Demiurge

Youtube Link: link

Youtube Channel Name: Gibson Verkuil

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

Demiurge

Youtube Title: Colloquium: Predictive and generative artificial intelligence for biology

Youtube Link: link

Youtube Channel Name: Illinois ECE

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

Colloquium: Predictive and generative artificial intelligence for biology

Youtube Title: The Quantum Physicist as Causal Detective: public lecture webcast trailer

Youtube Link: link

Youtube Channel Name: Perimeter Institute for Theoretical Physics

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

The Quantum Physicist as Causal Detective: public lecture webcast trailer

Youtube Title: The University Laboratory and Medical Device Translation: Successes, Caveats, and Pitfalls

Youtube Link: link

Youtube Channel Name: Fralin Biomedical Research Institute

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

The University Laboratory and Medical Device Translation: Successes, Caveats, and Pitfalls

Youtube Title: Rob A. Rutenbar: Taking Computing + Data Wide Across the Curriculum

Youtube Link: link

Youtube Channel Name: Berkeley EECS

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

Rob A. Rutenbar: Taking Computing + Data Wide Across the Curriculum

Youtube Title: Admin Law 2020 | Session 1: The New Landscape of Administrative Adjudication

Youtube Link: link

Youtube Channel Name: Duke University School of Law

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

Admin Law 2020 | Session 1: The New Landscape of Administrative Adjudication

Youtube Title: Rubbery Electronics Toward a Seamless Integration with Humans (Seminar)

Youtube Link: link

Youtube Channel Name: Dartmouth Engineering

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

Rubbery Electronics Toward a Seamless Integration with Humans (Seminar)

Youtube Videos

Youtube Title: Video1

Youtube Link: link

Youtube Channel Name: RobertVerkuil

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

Video1

Youtube Title: Intra-Body Communicator

Youtube Link: link

Youtube Channel Name: RobertVerkuil

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

Intra-Body Communicator

Youtube Title: Robo See Robo Do

Youtube Link: link

Youtube Channel Name: RobertVerkuil

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

Robo See Robo Do

Youtube Title: Video2

Youtube Link: link

Youtube Channel Name: RobertVerkuil

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

Video2

Youtube Title: Local Professor Gives Testimony at Trump Trial

Youtube Link: link

Youtube Channel Name: WLFITV

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

Local Professor Gives Testimony at Trump Trial

Youtube Title: Zandvoort Onboard 9-10-2020 1:499

Youtube Link: link

Youtube Channel Name: Thomas Verkuijl

Youtube Channel Link: https://www.youtube.com/@BMWE36Thomas

Zandvoort Onboard 9-10-2020 1:499

Youtube Title: Interoperability: A Foundation for the Future of Education

Youtube Link: link

Youtube Channel Name: 1EdTech

Youtube Channel Link: https://www.youtube.com/@1EdTech

Interoperability: A Foundation for the Future of Education

Youtube Title: Robert Young - Metric Embeddings and the Geometry of the Heisenberg Group

Youtube Link: link

Youtube Channel Name: Plectics

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

Robert Young - Metric Embeddings and the Geometry of the Heisenberg Group

Youtube Title: Linus Torvalds Guided Tour of His Home Office

Youtube Link: link

Youtube Channel Name: The Linux Foundation

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

Linus Torvalds Guided Tour of His Home Office

Youtube Title: Demiurge

Youtube Link: link

Youtube Channel Name: Gibson Verkuil

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

Demiurge

Youtube Title: Colloquium: Predictive and generative artificial intelligence for biology

Youtube Link: link

Youtube Channel Name: Illinois ECE

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

Colloquium: Predictive and generative artificial intelligence for biology

Youtube Title: The Quantum Physicist as Causal Detective: public lecture webcast trailer

Youtube Link: link

Youtube Channel Name: Perimeter Institute for Theoretical Physics

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

The Quantum Physicist as Causal Detective: public lecture webcast trailer

Youtube Title: The University Laboratory and Medical Device Translation: Successes, Caveats, and Pitfalls

Youtube Link: link

Youtube Channel Name: Fralin Biomedical Research Institute

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

The University Laboratory and Medical Device Translation: Successes, Caveats, and Pitfalls

Youtube Title: Rob A. Rutenbar: Taking Computing + Data Wide Across the Curriculum

Youtube Link: link

Youtube Channel Name: Berkeley EECS

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

Rob A. Rutenbar: Taking Computing + Data Wide Across the Curriculum

Youtube Title: Admin Law 2020 | Session 1: The New Landscape of Administrative Adjudication

Youtube Link: link

Youtube Channel Name: Duke University School of Law

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

Admin Law 2020 | Session 1: The New Landscape of Administrative Adjudication

Youtube Title: Rubbery Electronics Toward a Seamless Integration with Humans (Seminar)

Youtube Link: link

Youtube Channel Name: Dartmouth Engineering

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

Rubbery Electronics Toward a Seamless Integration with Humans (Seminar)