Working in the internal-spin up of a S&T program in AI/ML. Using advanced algorithms and ML/AI tools to do some hard and useful things with image processing for contract customers. Working to convert internally developed tools and technologies into viable defense solutions.

- Enabled tracking and alerting for USDA Cold-Treatment processes in refrigerated shipping containers, including updating production toolchain to be capable with advanced machine learning libraries (h2o.ai, ranger)- Established baseline meta-method for creation of FMA/FDD methods in refrigerated shipping containers- Assisted with efforts in false positive alarm reduction, automated D-optimal DOE design, and visual metrology serving other groups within the company.- Assisted with design of experiment and post-analysis for monitoring in new product development for large restaurant chains.- Coached/mentored/enabled larger team for Machine Learning, and automated disposition efforts.

• Apply machine learning to cybersecurity challenges in electromagnetic signal classification, adversarial robustness of image feature detection, and end-to-end toolchain development• Apply and evaluate forms of attentions (spatial, self, feature) in domain-to-visual transfer-classification to realize desired accuracy levels on specialized tasks.• Apply mechanical engineering and machine learning to mission identification and cognitive acceleration of reverse engineering to counter-adversarial UAS/UAV tasks• Extensive analog signal data analysis for classification tasks

My work provided critical and actionable understanding for work being done by several significant internal and external customers. I have also assisted internal groups with particularly challenging and informative data processing and interpretation. Given a pile of obtuse/sparse/grossly summarized data was able to find critical parameter that led to actual issue, and "run to ground". Proved the root cause was numeric and not physical. Built iSAX-informed machine learning tool to "reconstitute" system runs. Used reduced Koch's kappa with genetic-algorithm traversal of candidate formulae, and resulted in clear symbolic expression that perfectly binned the thousands of runs on all systems. Economic parameter value analysis on the expressions directly lead to standalone terms and simple experiments that enabled reproduction of the phenomena and reliable and repeatable reproduction of the weak/intermittent issue on all candidate systems. Built an automated test system to traverse the domain and characterize likelihood of event given driving parameters. Detailed data allowed adjustment of code that reduced rate of event to zero, retested and certified elimination of phenomena.Educated peers with basics of statistical process control for production, and to allow ongoing model learning. Made example D-optimal DOE that used substantially fewer points to achieve the same parameter uncertainty. Worked with internal data analytics team in creation, bench-marking, and tuning/polishing several high-value applications of machine learning.Novel thermal system innovation to improve SEER.

As part of startup process, trained the software tools (LabVIEW, etc), the team, the business model, the information architecture and the extended framework. Built and maintained assigned fault detection, diagnosis, and disposition (FDD) software that worked within existing and idiosyncratic framework. Demonstrated authoritative results with use of CART models and information criteria to validate heuristic mode segregation in alert preprocessing. Made a highly informative climate model for US that accounts for central and variational tendency in space, time and weather as well as resolving issues in typical use models with regards spatial under-sampling and multi-scale eigen-phenomena. Built internal-use only web-deployed, smart-device facing data access/analysis tools. Extensive data analysis using JMP, R, LabVIEW, MatLab, or Excel. Go-to guy for technically challenging database pulls and data analysis. Took on majority of engineering load for another engineer including root-causing misses, communicating to stakeholders, and polishing capability without negative impact to the uptime or integrity of his existing (large) code-base. Deconstructed, documented, and communicated the actual methods behind the cooling-mode code to allow reproducibility and extension of capability. Re-formulated our data format to SQLite3 in LabVIEW so it had ~2% of the memory footprint for the same information. Implemented compelling scoreboard (4dx) that clearly identified and documented algorithm health as separated from business driven policies. Created and proliferated single generic form for residential system fault-flags that allows plug-and-play alert code creation using a single settings file. Built software test framework using Caraya (jki) to certify algorithm/settings candidates against curated library of known and dispositioned cases.

ATTD - Visual Inspection Module - Module EngineeringCreated recipes, metrics, metrology methods, and performed light structural and system analysis in support of automated visual metrology systems. Worked in clean-room environment.Designed and executed experiments to determine underlying nature of physical systems as foundations for metrics and metrology. Ideated, created, tested/validated image processing techniques for detection. and classification of defects. Used robust statistics to create optimal, robust “golden unit” creation methods.Used finite mixture models and likelihood criteria in image processing to optimally identify material composition by region, particularly epoxy. Used feature-let analog of wavelet method to perform very high accuracy sub-pixel feature measurement from video footage to support defect detection at spatial scales below the wavelength of illumination and beyond the imaging resolution of the camera.Automated DOE (design of experiment) process on tool to improve data capture rate by 18x, and allow significant improvement in quality of inspection and classification of parts.Performed data finding, preprocessing, analysis, and conclusion validation for a number of stakeholders in a variety of adjacent business units.Used D-optimal DOE to design illumination caliber for spatial and intensity matching.Found and prototyped non-parametric formulation of coplanarity for capture of customer-facing defects.Assisted a number of platforms/products in effective recipe creation.Staged effective document management including creation and training on documentation tools, BKM mapping and management, and staging data-transfer and certification templates for internal customers.

Provided high-level structural dynamic models, simulation and results to inform design for shock mitigation and ensure solder-joint reliability over typical use conditions. Characterized Risk by simulating system under statistically derived load cases and conditions, characterizing failure envelope, processing data and results, summarizing and presenting results, and making recommendations. Characterized failure envelopes of component chips, solderballs, and other materials – enabling mitigating structural design and design optimization. Designed test boards and test requirements for memory chip SJR evaluation.Implemented and validated automated material parameter tuning that coupled FEM simulation parameters to empirical data – creating a foundation for optimal model calibration. Determined optimal preload values for loaded structures under shock. Validated results with shock test and dye/pry results.Made preliminary (first order) PID controller for fan speed control using MatLab control toolbox to support development efforts in neighboring groups.

Used SPC on provided utilities to find and resolve OOC issues on daily basis. Performed monthly and quarterly trend reviews to adjust control settings for optimal control. Extracted data from SQL-server database using Transact-SQL, analyzed with JMP software and Excel, evaluated and validated results and recommended responses for wide variety of customer issues. Gathered data, constructed equipment schedules, built connectivity diagrams, characterized capacity and reliability as part of site-wide risk assessment, capacity, and energy efficiency analyses. Researched mechanical components and control strategies for capital and recurring cost and presented results to management.

Gathered data, and used MatLab and system identification techniques to noninvasively characterize mechanical properties of biological systems. System Identification and Parameter Estimation using Recursive Least Squares (RLS), Kalman Filter (KF), Extended Kalman Filter (EKF) and Particle Filter applied to mechanical parameter estimation in cardiac tissue. Operated and troubleshot LaVision DPIV (Particle Image Velocimetry) system and DaVis software and extracted numeric samples of 3d dynamic fluid flow-fields. Performed fluid flow analysis including ensemble mean, and characteristics. Interacted with Mayo clinic to gather specifications, design, construct, and operate novel sonogram test cell as part of heart valve diagnosis research. Lab expert in MatLab.

Gathered data, and used MatLab and system identification techniques to noninvasively characterize mechanical properties of biological systems. System Identification and Parameter Estimation using Recursive Least Squares (RLS), Kalman Filter (KF), Extended Kalman Filter (EKF) and Particle Filter applied to mechanical parameter estimation in cardiac tissue. Operated and troubleshot LaVision DPIV (Particle Image Velocimetry) system and DaVis software and extracted numeric samples of 3d dynamic fluid flow-fields. Performed fluid flow analysis including ensemble mean, and characteristics. Interacted with Mayo clinic to gather specifications, design, construct, and operate novel sonogram test cell as part of heart valve diagnosis research. Lab expert in MatLab.

Mechanical Engineer for Multidisciplinary, Integrated Product development Team Sponsored by Intel HealthcareDesigned Advanced Hearing Aide System Technology. Idea contributed significantly to headphone and hearing aide design in the next 2 years including making noise-cancellation headphones mass-producible, and noise cancellation and voice clarifying hearing aides.Participated in and contributed to every phase of product development including understanding the market, ideating products, selecting a subset to develop, refining the product ideas, soliciting and responding to final product selection, designing the production, marketing, and performance of the product, and optimizing manufacturing, social and environmental costs.Update: This was completed in 2007, but in April 2014 I interviewed with Michael Nikkhoo (michael-nikkhoo-13a5014), a director at a stealth hardware group for Microsoft in Mountain View, who shortly left to return to work at Apple and is now a senior technologist and engineering director. I sent him a copy of this content to show him my work. Almost exactly one year later Apple released the Apple watch. Now look at the earbud on slide 15 of the final innovation proposal and compare it to the Apple AirPod.

Worked with advanced missile control system component production.

Worked to reduce noise emitted by air-compressor enclosure, and evaluated thermal systems to reduce internal operating temperature issues. Directly correlated noise characteristics to geometric design allowing design improvement to change emitted noise spectrum. Improved product reliability, marketability, and manufacturability. Contract and Part-time position during school.

Worked on using Gridgen to build a CFD grid for the dimpling of a golf-ball.

Established tooling and processes, managed production and employees, and troubleshot equipment to resolve issues as owner of decal golf-club production line. Limited role as Network and System Administrator. (This was a summer job during college)

tutored a wide variety of students on subjects ranging from algebra through calculus III, physics, chemistry, statics, dynamics, and numeric methods.

Moved 3 tons of freight a night into the store and put it on shelves. It was a summer job but it wouldn't work with class schedule for fall courses at MCC so I had to quit and keep working on college. Was able to get job tutoring math.

Designed flashlight parts including: head, pack, heat sink, o-rings and assembly tools. Developed assembly methods, trained workers. Increased production by 300%. Moved technology in-house. After working in this position I started attending Mesa Community college and working on my Bachelors degree.

Worked under direct instruction of Director of Technology for campus. Researched software, components and strategies. BUILT software tools. Engineered a campus-wide video projection & internet infrastructure upgrade. Served as departmental expert in Excel and PowerPoint. Certified in Cisco Network Administration.

Assisted in design of hi-low driver circuitry, z-axis carriage head, and hardware control interface. Performed price and availability research for effective purchase of components and processes including cable harness, electronic components, materials, fabrication, powder coating, and shipping.Performed detailed research for the purposes of component selection of 8 and 16 bit microcontrollers, hand plasma-torches, belt drives, and parallel port standards. Provided Phoenix-Area on-site technical support to customers.Primary software tools were Solidworks, PADS, and Microsoft office.

Worked under NDA to design/test safety equipment for Apache Helicopter. Performed FEA analysis and optimization of pyrotechnic housing using Solidworks and COSMOS/M. Assisted as part of team in manufacturing, testing, and evaluation of pyrotechnic pre-tensioner components. Researched and appropriately applied Mil-Specs, and appropriately selected high-performance gear steels.Performed Circuit and PCB layout using PADS PowerLogic and PowerPCB. Assembled parts of G-tester devices. Received Excel, and Solidworks training.