Problem: A new roof design called for a unique heat-cure bonding process to achieve part quality and cycle time requirements; there was no specification or approved material for this application.
Solution: I surveyed candidate adhesives from (6) different suppliers, performed key screening tests to downselect options, and developed a test method to simulate the unique cure process. I collaborated with Engineering to define the specification for the adhesive, validated the preferred product according to the specification, and audited prototype and early production parts by teardown to confirm bonding process quality.
Outcomes: An adhesive compatible with the desired process was identified and approved. Material requirements and test methods for determining material performance were standardized for management of change and use in future part designs. Prototype parts meeting the new requirements were available for critical tests to keep the program on schedule.
Business value: Supported launch of $900M product line.
Problem: Variability in supplier quality of sealant applied to the (4) highest volume cab models led to application defects (intermittent, up to 3% defect rate) at drum change-out (once per shift) and required additional operator time to adjust the application parameters and rework cabs.
Solution: I engaged the incumbent supplier to improve quality and engaged (2) new suppliers to provide an alternative solution. Incumbent was not able to improve quality. I managed technical relationships with the prospective suppliers by communicating product requirements and guiding them through key screening tests to generate a preliminary product data package. I reviewed and vetted supplier data by reproducing key tests in-house and at third-party test labs. I then planned and executed validation of the preferred supplier’s sealant according to company standards and verified key performance requirements (sealant adhesion and paint adhesion) of the new sealant on prototype cab builds. I collaborated with Manufacturing Engineering to deploy a safe-launch process for sealant quality while the new sealant was launched.
Outcomes: A new sealant was approved for use and implemented in the factory. Application defects were eliminated enabled production of $125-250k worth of additional product per day.
Business value: Eliminated 45-minutes rework per shift. Increased plant output by $350K finished product per day.
Problem: Procedures for documenting bonding process quality of body panels was not standardized and was not practical for tier suppliers to perform on a routine basis. The supplier’s procedures were inadequate. Bonding process defects could escape detection for long periods of time.
Solution: I integrated requirements from all joining and sealing standards (adhesives, sealants, tapes, rivets, hems, welds) into a comprehensive analysis procedure for components. I defined methods for documenting and quantifying bond and joint quality using readily available tools. I developed a workflow for tabulating and scoring component quality and participated in a pilot trial of the standardized process with our largest tier supplier.
Outcomes: A procedure for documenting and quantifying bonding process quality was standardized and gave Engineering, Quality, and Tier suppliers a common framework to discuss bond quality. Supplier engagement and feedback was positive and identified aspects of the quantification process that could be streamlined given supplier automated vision system capability and areas that could be simplified to aid implementation. Quality had a tool to use in reducing the pool of potentially defective parts.
Business value: Reduced warranty risk pool from $8B to $640M.
Problem: Windshield moldings on the (2) highest volume models exhibited severe blistering and degradation after system checkout; defect started intermittently and rose to a >75% failure rate. This resulted in a large quarantine pool, large rework queue, and delayed deliveries of finished product.
Solution: I inspected defective parts (qualitative flexibility, microscope, FTIR) and used these observations to identify potential root-causes with Engineering, Quality, Tier 1 and Tier 2 suppliers. I developed a test method to reproduce the failure and deployed this test method to Quality teams to audit inventory and product in the field. The root cause was supplier equipment failure leading to improperly cured moldings.
Outcomes: A rapid test procedure was developed to inspect parts at all stages of the supply chain and prevent installation of defective parts into our products while the supplier resolved the root cause. Observations made in the lab narrowed down possible root causes to help focus the response team on evaluating those related to the molding process.
Business value: Resolved an issue affecting $40M of product
Problem: The company did not have documented specifications for foam tapes; engineers would select products based on prior experience or supplier recommendations and did not have standard methods for validating performance or for identifying alternative products when necessary. This led to confusion among design engineers and line operators about which tapes could and should be used in different applications. Lack of acceptance criteria made rejecting suspected defective material nearly impossible.
Solution: I identified the (4) different use cases for foam tapes across (6) product lines at (3) different factories. I authored the standard defining the specifications and test methods for each use case, reviewed this standard with Engineering for approval and validated the performance of the approved products for each use case.
Outcomes: Engineering drawing callouts for foam tapes were made consistent with company best practices and reduced the risk of incorrect application. Material requirements and test methods for determining material performance were standardized for management of change, use in future part designs, and quality control.
Business value: Standardized tape products used in badging, reinforcement, sealing, and assembly on vehicles with build volumes over >60K units per year and $8B in revenue.
Problem: Automotive OEMs and tier suppliers needed a bonding process that eliminated quality issues from application of liquid adhesives, was easy to apply, and was fast curing to eliminate work-in-process.
Solution: Led product development of an epoxy/acrylic adhesive with adhesion to glass, metals, and plastics. The adhesive could be supplied in a film or tape form for easy application, cured upon exposure to UV-light, and had strong initial holding power to eliminate need for fixturing during cure. I led cross-discipline collaboration efforts between Engineering, Corporate R&D, Sales and Marketing, and Regulatory to optimize the UV-curing process, implement a proprietary multi-layer coating process, engage customers, and ensure product compliance.
Outcomes: Led the team through three rounds of MVP (minimum viable product) feedback cycles. One iteration was included in an OEM vehicle prototype build for technology scouting; the prototype survived extensive durability testing. This work resulted in (6) granted patents to protect our competitive edge. The project continued after I transitioned off the team; the commercialization team focused on manufacturing process scale-up, application engineering, and “spec-in” at the big three US automotive OEMs and their tier suppliers.
Business value: Drive organic business growth by commercializing new products ($110M opportunity).
Problem: Automotive OEMs needed alternatives to existing products and there was expected growth in this market. I led development of the general-purpose product in the portfolio for hem flange bonding.
Solution: I “de-tuned” a high-performance product to give a general purpose product formula meeting cost, performance, and shelf-life requirements . Through in-house testing, I validated that the prototype formula met (3) different customer specifications. I collaborated with Sales and Marketing and Regulatory to prepare product sales literature and regulatory documentation. I facilitated pilot manufacturing trials and collaborated with Counsel to file a patent application on the formula.
Outcomes: Customers gave positive feedback about the performance; one customer agreed to perform independent specification testing as the first step toward acceptance. Business priorities changed and the project was shelved.
Business value: Drive organic business growth by commercializing new products ($10M opportunity).
Problem: Customers needed an antifungal finish for semi-finished goods that was less toxic than their current product.
Solution: I developed two non-toxic chitosan-based finishes compliant with EPA minimum risk pesticide requirements that also met the customer's specification for antifungal performance in third-party laboratory testing. I led a cross-functional collaboration with Operations to complete pilot-scale trials and define quality and release testing.
Outcomes: Test data and product samples were shared with the customer for feedback and secured a line-trial at the customer.
Business value: Drive organic business growth by commercializing new products ($1.4M opportunity)
Problem: Requests for prototypes of custom products from prospective customers took 4+ weeks to fill and requirements were often not collected.
Solution: I developed and implemented a rapid prototyping system delivering 3-15 prototype polymer formulations per week for business development. Collaborated with business development, sales, manufacturing, regulatory, and fulfillment to define system requirements and iterate toward the implemented solution in 5 months. Built database of existing product and experimental product formulations with recipes, properties, and performance. Built intake process to systematize requirements collection. Hired and trained a technician to plan, produce, and characterize prototypes. Built an information radiator to show request status, sample details, and customer feedback.
Outcomes: Reduced the lead time for prototype ideation-to-delivery by 75%, achieved 95% user adoption rate and achieved 90% product documentation rate.
Business value: System kept prospective customers engaged in the validation of new product ideas and organized product data for repeat orders and product refinement.
Problem: Early stage startup needed to quickly identify and engage analytical labs capable of performing specialized polymer analysis to support process and product development.
Solution: I engaged 8 third party analytical labs to determine capabilities, capacity, and cost. I identified and vetted preferred service providers within 4 weeks and I gathered and organized key source selection criteria for an objective selection process to meet the startups requirements.
Outcomes: Expanded startup's access to analysis necessary for product development and customer engagement to find product-market fit.
Business value: Accelerated startup's selection of testing labs by schedule crashing. Allowed startup to initiate product analysis on schedule.
Problem: Lift tickets are expensive! Tracking the weather and choosing which ski area to visit for the best experience can be time consuming.
Solution: Use this page to compare daily NOAA forecasts for seven major ski areas in western Washington.
The app is built using the following tools:
Python, HTML, CSS
Azure Functions
Azure Blob Storage
This app collects forecast grid data once per day from the NOAA weather API for each ski area and saves the raw data as JSON in a blob storage container.
Logic is included to handle HTTP request errors and changing NOAA endpoints.
After successfully retrieving forecast data for each location, the app loads each forecast from blob storage and creates a forecast object. The forecast data are parsed and grouped by day. Next, depending on the weather property, the maximum, minimum, average, and total values are calculated for the AM, PM, overnight (ON), and for the full 24 hour period from 6am to the following 6am. Then, measures are converted from SI to US Customary units. Finally, the status for each day is determined and set by the above criteria.
Logic is included to handle null property data and manage the variable frequency of forecast timepoints.
Forecast data for each day are scored according to the criteria described here.
Data for each location and each day are then used to build cells in a table object. The table is then written to HTML. The HTML file is stored in a blob storage container configured for web access.
Objective: Analyze US Census Bureau County Business Patterns data and US EPA Pollution Factors data to identify industries and locations to target for policy change.
Methods: Raw data was retrieved from .csv files or as JSON from API calls. Data was cleaned and analyzed using the Python libraries Pandas and NumPy. Feature engineering involved calculating average annual wages per employee and pollution factors by GDP codes.
Visualization was accomplished using Seaborn and Matplotlib. See the notebook here.
Conclusions: The main conclusions drawn from this analysis are:
Top polluting industries (kg/USD) were Waste Management (1.9), Pipeline Transportation (1.6), and Farms (1.5)
States with the largest growth in these industries were NM (+50%), NH (+1000%), and IL (+80%)
Of the top ten polluting industries, Truck Transportation, Mining, and Non-metallic Mineral Manufacturing had the highest wage growth (15-16% yoy)
Objective: The Ford GoBike dataset contains >183,000 telemetry entries for eBikes used in San Francisco during February 2019. Data includes information about ride start and end times, ride duration, rider gender and age, ride start and end location, and member status. This dataset was analyzed to identify rider use patterns, determine peak use times, and typical ride duration.
Methods: Exploratory data analysis was performed using the Python libraries Pandas and NumPy. Visualization was accomplished using Seaborn and Matplotlib. See the notebook here.
Conclusions: The main conclusions drawn from this analysis are:
GoBikes are used primarily for short trips (average = 11m 45s)
Peak weekday use coincides with typical business commute times
Weekend use is less than weekday use
Weekend use peaks mid-day and ride duration is slightly longer than weekdays
21 - 40 year olds are the most frequent users
Use patterns are similar for all genders and age groups
Objective: NICS Firearm Background Check and US Census data were evaluated to identify possible correlations between demographics and firearm background checks. Background checks serve as a proxy for gun sales.
Methods: Raw data was loaded from .csv files. Data were cleaned, analyzed, and visualizations were created all using the Python library Pandas. See the notebook here.
Conclusions: The main conclusions drawn from this analysis are:
Texas (88k/mo), California (71k/mo), and Florida (66k/mo) have the largest average number of monthly background checks while low population density states such as Alaska (8.0e-3), West Virginia (7.8e-3), and North Dakota (7.7e-3) have the largest per-capita average monthly background checks
For the six-year period ending in 2016, Illinois (+234%), Massachusetts (+229), and New Jersey (+226) had the largest increases in background checks
There is a positive relationship between increases in background checks and economic factors such as per-capita income, median home values, median rents, number of building permits, and rates of college education