Thomas Eicher Email & Phone Number

Goodyear, Arizona

General software development and engineering support over full life cycle. Compute intensive scientific, signal and image processing applications under Linux using C/C++, Matlab and scripting languages.

Palmdale, California (Skunkworks)

Duties: Developed and maintained embedded real-time software using C language under VxWorks/Tornado environment. Maintenance of requirements and design documents.Accomplishments:Analyzed and corrected problems involving avionics hardware in the loop caused by misinterpretation of vendor ICD and/or operational requirements. This resulted in greater customer satisfaction when long standing defects were resolved. Team effort in this area was recognized by cash award.Developed software based monitors to record and retrieve hardware operational metrics for non-volatile memory, including on-off cycles and power on time. This provided a source of data used for preventive maintenance, resulting in fewer breakdowns in the field. Conducted and participated in engineering reviews and inspections to ensure compliance with specifications and requirements. This eliminated many defects at the source and resulted in a reduction of problems encountered in the field.

Goodyear, Arizona

Duties:Team development of high performance scientific algorithms and associated software implementations under Unix and Linux. Functioned as liaison between specialist algorithm design teams (scientists and mathematicians) and software implementation teams (software engineers). Accomplishments:Addressed critical system testing requirements by collaborating with senior scientists to develop high fidelity simulation. The simulator was written under Unix using a combination of Matlab, C and Fortran and produced ICD compliant input data for a large scale data processing system. The data produced by this simulator facilitated algorithm development and allowed the processing system to be fully verified, validated and debugged prior to delivery.Successfully implemented mathematically complex parallel processing algorithms under a tight schedule. Implementations were completed on schedule, allowing the overall project to meet schedule.Analyzed and resolved a complex problem involving custom math accelerator hardware that resulted in complete scrambling of output data. This was a critical issue, with visibility to the level of corporate vice presidents. The root cause was eventually determined to be incorrect memory alignment related to the custom hardware using direct memory access.Analyzed execution timeline of large parallel processing applications running on SGI NUMA systems to identify performance bottlenecks. Subsequently devised solutions to improve execution timeline, such as integrating optimized numerical libraries (e.g. Intel MKL) or incorporating multi-threading (OpenMP).Led efforts to evaluate and benchmark HPC systems. Devised benchmarks to test floating point processing rates, CPU memory bandwidth and sustained I/O rates. Coordinated with vendors and network administrators to ensure timely availability of test environments. Produced reports detailing benchmarking results.