Joseph Carmack Email & Phone Number

Merrimack, New Hampshire, United States

My work at BAE Systems is in machine learning and artificial intelligence applied to the RF domain.I've worked as the algorithm lead on several DARPA RF Machine Learning programs (e.g. DARPA RFMLS and DARPA SPiNN) to develop deep learning technologies for the RF domain that include new state-of-the-art RF Fingerprinting deep classification, deep waveform synthesis, multi-agent deep reinforcement learning for RF waveform manipulation, end-to-end LTE receiver with deep learning based signal processing kernels for interference cancellation, time and frequency offset correction, and multipath propagation channel equalization.I've been the PI for multiple internal R&D efforts to develop explainable AI for the the RF domain and enhance the robustness of RF Fingerprinting technologies using architectural inductive bias. Other significant projects include developing a framework for doing fast prototyping of deep reinforcement learning algorithms on new custom environments and then scaling those prototypes to massively parallel implementations to achieve optimal results.Much of this work has resulted in peer reviewed published research articles, conference presentations, and intellectual property. See my google scholar profile for a list of publicly released papers. There are also 4 pending patents for which I am an inventor.

Fayetteville, Arkansas Area

I am pursuing a PhD degree under Dr. Paul C. Millett in computational materials science. My PhD research is focused on studying the self-assembly of nanoparticles suspended in immiscible polymer blends and their potential application to create thin flexible membranes with nanoparticle lined channels for use in exploiting the nanoparticles superior catalytic properties with applications in CO2 remediation, energy storage and conversion, and filtration.I use the phasefield method coupled with brownian dynamics (very similar to molecular dynamics) for my current computational modeling at the mesoscale. All our computer models are run on massive supercomputers which utilize MPI to distribute the computational work across hundreds of processors.Using computers to do science has become a passion of mine. I am currently developing my interests in machine learning and artificial intelligence with the hope of using these skills to do better science and engineering. I have learned to program GPUs, the computational work horse responsible for training machine learning algorithms. GPUs are also very capable of accelerating existing scientific codes which is another area of development for me.

Idaho Falls, Idaho

I worked with the fuels modeling and simulation group. Our job was to improve the fuel performance of nuclear fuel. Through high fidelity simulations performed on supercomputers we studied the complex coupled physics going on in nuclear fuel in the conditions of a nuclear reactor core. I gave an oral presentation describing the research results I obtained under the mentoring of Dr. Yongfeng Zhang at the fall 2013 Minerals, Metals, and Materials Society (TMS) meeting.

Rexburg, Idaho

I taught the lecture of a 3 credit physics course. I had to organize the course, assign and grade homework, and meet with students who needed help.

Provo, Utah Area

NSF undergraduate research internship. I worked with physicist Dr. Branton Campbell in extendeding a new data analysis technique to determine the magnetic structure of materials from experimental x-ray and neutron diffraction data.

Doctor Of Mechanical Engineering (Phd), Computational Material Science And Engineering, 4.0 Gpa

I will be working under Dr. Paul Millet, a new faculty member in the department whom I met at Idaho National Lab in the fuels modeling.

Bachelor Of Science (Bs), Major: Physics Minor: Mathematics, 3.86 Gpa

Activities and Societies: Sigma Pi Sigma, Society of Physics Students, American Physical SocietyAdvance physics course work in solid-state.