Adam Thompson Email & Phone Number

Alexandria, Va, Us

Santa Clara, Ca, Us

As a Principal Technical Product Manager, Adam is responsible for defining, building, and evangelizing NVIDIA's software and hardware platforms targeted towards real time sensor processing for computational instruments.Computational instruments are the cornerstone of scientific advancements: radio telescopes search the heavens for origins of the universe and quest for intelligent lifeforms; light sources emit high speed x-rays to discover how materials behave and interact with others; software defined radios enable efficient and reliable communication across complex environments.With the promise of AI analytics at the edge, intelligent sensors, and integrated edge-to-datacenter/cloud collaboration - including on-line AI model retraining and sensor-defined digital twins, it's imperative to deliver a GPU-based sensor processing platform that's performant, scalable, and easy to use.One product in his portfolio is the Holoscan SDK, a domain-agnostic and AI-enabled sensor processing library to build and deploy high performance pipelines. With Holoscan, a developer can quickly connect high bandwidth, low latency sensor data to GPU and seamlessly build application pipelines with NVIDIA software in both Python and C++. Trained AI models can be delivered to this real time pipeline with ease, meaning a developer can go from a model-zoo download to real-time inferencing with little time investment.If you're looking to revolutionize your sensor processing pipelines with NVIDIA GPUs, please reach out!

Santa Clara, Ca, Us

Santa Clara, Ca, Us

As a Solutions Architect, Adam was a customer-facing engineer who helped our customers deploy GPU-based solutions, resulting in faster and better performance for their specific use cases. At the same time, he served as an advocate for those customers internally by identifying current gaps in software, hardware, or networking coverage and working closely with engineering to improve the overall developer experience.His specific area of technical focus at NVIDIA has been with communications, sensors, and signal processing. Real time, high bandwidth, and low latency applications have long been dominated by accelerators like ASICs and FPGAs, but with the flexibility, performance, and programmability of the CUDA software stack, in addition to the desire to apply ML/DL based techniques to time-series data, GPUs make an attractive alternative. With a strong focus on the intersection of both 5G/6G communication systems and the needs of Public Sector customers (radar, sonar, telecom, security, etc), he has helped build the overall signal/sensor GPU platform - from software-defined radio and networking through AI inferencing.A core piece of this work is the cuSignal software library. He created cuSignal in 2019 to serve as a GPU-accelerated Python library targeted towards elemental signal processing functions (convolution, correlation, spectrum estimation, filtering, etc). It extends the SciPy Signal API and is widely used in both the commercial and public sector industries. cuSignal has been downloaded over 400,000 times via Anaconda alone and is the defacto software library for fast signal processing and AI development. As of CuPy v13 (released in January of 2024), cuSignal has been deprecated as an NVIDIA product and is now available in CuPy.

Washington, Dc, Us

Fall 2021 COSC 155 Course: GPU Computing with Python

Us

Systems engineering, development, deep learning, and program management targeting defense applications dependent on a larger collection framework.Responsible for government and internal communication, project design, metrics, application performance, contract deliverables, and financial status.

San Diego, California, Us

Refine and improve signal processing algorithms by focusing on computational cost and execution time - using MATLAB mex files, C/C++ code, and GPU processing when necessary.

Reston, Virginia, Us

Develop 1-D and 2-D model based algorithms for the lossy and lossless compression of complex and detected imagery over various sensor types.Present algorithm status updates and documentation to both technical and non-technical customers.Design high performance computing trade studies to determine the feasibility of running various signal processing algorithms on a computer’s graphics card.

Atlanta, Georgia , Us

Continued support and development of GPU-VSIPL. This library allowed pre-existing Core-VSIPL code to be run on a computer's GPU - in effect achieving significant run-time speedups without exposing the programmer to oftentimes obfuscated and confusing CUDA code. For this project, I was responsible for:* Fixing reported bugs* Complex and real toeplitz general system solver* Complex and real linear least squares solver

Reston, Va, Us

Designed and implemented an automatic, real time, detection and classification system for identifying synthetic speech signals occuring within the HF band.Wrote a CUDA program to estimate the power spectral density of a signal, specifically the output of a polyphase channelizer, using the Welch periodogram. By running the program on an NVIDIA graphics card, computational speed was increased by an order of magnitude.

Tutored students in Signals, Systems & Transforms, Basic Circuits, and Electrical Engineering for non-electrical engineers.

Huntsville, Al, Us

Analog Group:Designed various buck converters to be used in power supplies.Created circuits through DX Designer, laid out boards with Allegro, and manually placed circuit elements.Total Access System Core Software:Designed a Tcl automation test suite to facilitate the testing of voice and IP data across an 8 node multi-chained Total Access 5000 system, resulting in the reduction of testing time. This script was later implemented in both the corporate and field offices.Served as the ADTRAN student ambassador to Clemson University.DSL Development - Design Verification Testing:Responsible for the complete testing of two customer code revisions for the Total Access 2nd Generation ATM based 1148 DSLAMS.

Ms, Electrical & Computer Engineering

Bs, Electrical Engineering