Jonathan Mckinney Email and Phone Number
Jonathan Mckinney work email
- Valid
- Valid
- Valid
Jonathan Mckinney personal email
- Valid
- Valid
15 years experience developing machine learning models, including AutoML frameworks and training and deploying large-language models.* #2 top LLM developer in world in 2023: https://huyenchip.com/llama-devs* Co-developed h2oGPT at h2o.ai: https://github.com/h2oai/h2ogpt -- Large Language training and generation framework including chatbot.* Co-developed DriverlessAI at h2o.ai -- AutoML framework for multi-modal models (NLP, image, tabular, etc.)20 years of experience building world-class mathematical algorithms and astrophysical models, and executing them on massively parallel systems to explain and predict real-world data.* >100 peer-reviewed scientific publications in top journals, including Science* >10000 refereed citations with refereed h-index > 50 (http://bit.ly/1Pve92Y and http://bit.ly/1PAsiKR)* ~$3 million in grant funds as Principle Investigator, ~$9 million as co-PI* ~15 students advised (at Harvard/Stanford, now with prestigious fellowships)* ~20 graduate and undergraduate courses taught* ~70 invited conference talks and public lectures* ~4 international teams, as senior member, involving ~500 people* ~120 million cpu-hours of computing time awarded, worth ~$4 millionDesigned and Built: http://github.com/pseudotensor* Deep Learning Projects: CNNs and LSTMs to predict video, faster convergence for gradient descent, and new architectures. Using Tensorflow, Theano, and Matlab.Astrophysics Simulation Projects:* Built: (90% as designer and developer): HARMRAD/HARM: World-leading massively parallel general relativistic radiation plasma simulation and analysis codes (600k lines of code). Impact: used by >50 researchers in >100 articles.* Built: (95% as primary designer/developer): Many projects: fluid, radiation, optimization, nonlinear solving, signal processing, simulated annealing, Markov chain Monte Carlo, visualization. Impact: used by dozens of researchers in >100 articles.* Scaling: (as 100% designer/developer) Run on >100k cores, with petabytes of data* Languages/APIs used: C, Python, C++, Fortran, MPI, OpenMP, CUDA, OpenCL, OpenACC, Matlab/Mathematica and parallel debuggers (Intel Inspector)* Algorithm Projects: (as 100% designer/developer): tensor algebra solutions to general relativistic inversion problem, radiation/fluid numerical optimization, integro-PDE solvers
H2O.Ai
View-
Director Of ResearchH2O.Ai Mar 2017 - PresentMountain View, Ca, UsDeveloping h2oGPT (https://github.com/h2oai/h2ogpt), H2O.ai's large language model (LLM) chatbot with training, generation, and UI code.Developing Driverless AI (https://www.h2o.ai/driverless-ai/), H2O.ai's automatic model selection and tuning (e.g. parameters for Torch, Tensorflow, LightGBM, XGBoost via AutoML involving multi-modal data including tabular, image, text via NLP) and evolutionary-based feature-engineering product with control over interpretability.Developing H2O4GPU (https://github.com/h2oai/h2o4gpu), the next-generation h2o, with fast GPU and CPU machine learning algorithms.
-
Professor Of PhysicsUniversity Of Maryland Jul 2012 - Dec 2019College Park, Md, UsComputational Physics and Astrophysics, Theoretical Astrophysics, Machine Learning, Markov Chain Monte Carlo, GPU acceleration. Focusing on testing Einstein's general relativity theory using the Event Horizon Telescope and studying the cosmological formation of galaxies and black holes in the early universe. -
Visiting ScholarStanford University Dec 2016 - Dec 2018Stanford, Ca, UsResearch and Applications in Deep Learning (and general Machine Learning and Artificial Intelligence). Focusing on convolutional neural networks (CNNs), long-short term memory (LSTM) networks, new architectures, improved gradient decent (https://github.com/pseudotensor/GeoAcc), self-optimizing self-organizing networks, etc. as applied to text and video prediction (e.g.https://github.com/pseudotensor/temporal_autoencoder) . Primarily using Tensorflow, but also using Theano and Matlab. Working with Stefano Ermon (Stanford AI Lab). -
Nasa Einstein FellowStanford University Sep 2007 - 2012Stanford, Ca, UsComputational astrophysics to study black hole accretion disks, jets, and core-collapse supernovae. Using techniques like fluid dynamics, finite volume methods, machine learning. Prize fellowship. Worked with Roger Blandford (Stanford faculty and former director of KIPAC).
-
Guest LecturerStanford University Sep 2007 - 2008Stanford, Ca, Usundergraduate level, Black Holes, Prof. Tom Abel
-
Scidac (Scientific Discovery Through Advanced Computing) Postdoc FellowSlac National Accelerator Laboratory 2010 - 2011Menlo Park, California, UsMassively parallel computations using MPI, OpenMP, CUDA, to accelerate plasma simulations with physics solvers and optimizations (machine learning) -
Harvard Institute For Theory And Computation (Itc) Postdoc FellowHarvard University Sep 2004 - May 2007Cambridge, Massachusetts, UsComputational Astrophysics, focusing on new equations of motion and new optimization techniques with machine learning. Prize fellowship. Worked with Ramesh Narayan (Harvard astrophysics faculty). -
Nasa Graduate FellowUniversity Of Illinois At Urbana-Champaign Sep 1998 - Sep 2004Champaign, Il, UsAstrophysics of black holes and plasmas, focusing on testing accretion and jet theories using massively parallel simulations of accreting black holes. Award from NASA competition to fully pay for graduate research stipend. Worked with Charles Gammie and Stu Shapiro (UIUC physics and astronomy faculty).
Jonathan Mckinney Skills
Jonathan Mckinney Education Details
-
University Of Illinois Urbana-ChampaignPhysics -
Texas A&M UniversityPhysics
Frequently Asked Questions about Jonathan Mckinney
What company does Jonathan Mckinney work for?
Jonathan Mckinney works for H2o.ai
What is Jonathan Mckinney's role at the current company?
Jonathan Mckinney's current role is Director of Research at H2O.ai, Astrophysics Professor at UMD, Visiting Scholar at Stanford.
What is Jonathan Mckinney's email address?
Jonathan Mckinney's email address is jm****@****ord.edu
What schools did Jonathan Mckinney attend?
Jonathan Mckinney attended University Of Illinois Urbana-Champaign, Texas A&m University.
What are some of Jonathan Mckinney's interests?
Jonathan Mckinney has interest in Science.
What skills is Jonathan Mckinney known for?
Jonathan Mckinney has skills like Physics, Numerical Analysis, Science, Astrophysics, Scientific Computing, Theory, Research, Fortran, Latex, Mathematical Modeling, Plasma Physics, Simulations.
Free Chrome Extension
Find emails, phones & company data instantly
Aero Online
Your AI prospecting assistant
Select data to include:
0 records × $0.02 per record
Download 750 million emails and 100 million phone numbers
Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.
Start your free trial