Resumé
"My experience has been nothing short of exceptional on all levels, but primarily for two key reasons. First, the quality of programming has been so high that it offsets the presumed cost advantage of overseas development opportunities. Second, and probably most significant, I receive an uncommonly high level of service with regard to programming issues, customer concerns, and hosting challenges. I’ve been in various aspects of software business for over 20 years, and have had my own software business for over three years. Kevin Weller is as good as it gets when it comes to development and service. He gets my highest recommendation without qualification." — Mark Lipsitt, CEO, The Lipsitt Group
“Managed expectations, never overcommitted, and consistently delivered on time.” — Dewayne Korth, Senior Application Architect, BLUEFIN
Experienced and Reliable Software, Data, AI/ML and Integration Architect and Engineer
Instructor
Contact Information
Areas of Expertise
- Ensuring business value by listening and asking the right questions
- Multi-tier Web Application Development and Technical Leadership
- Artificial Intelligence (AI) Solutions Architecture and Implementation
- Advanced Geospatial Algorithm Design and Implementation
- Geographic Information Systems (GIS) Architecture and Buildout
- Technology Sales, Marketing, and Business Strategy
- Technology Solution Return On Investment (ROI)/Net Present Value (NPV)/Breakeven Analysis
- Market Research Support and ML/AI Analytics
- BootcampGIS.com Course Author and Instructor on GIS/CAD Integration using AWS and the FME Platform
- Enterprise Cloud and On-Premise Systems Integration and Migration
- Enterprise Business Process Automation using scripting languages, FME® Flow, ActiveBatch, etc.
- Automated Data Pipelines for Extraction, Validation, Cleansing, Transformation, Loading, and Analytics (ETL, ELT, event-driven, streaming) using FME®
- Reusable Modular Component Library Strategy, Development and Maintenance
Professional Certifications
- Certified FME® Professional
- Certified FME® Flow Professional
Technical Summary
Languages
- Python
- Rust
- Ruby
- JavaScript
- C#
- SQL
- HTML
- Shell scripts
- C
- Objective-C
- Speedware/4GL
- Natural/2
AI/ML
- OpenAI
- Anthropic
- scikit-learn
- LangChain
- PyTorch
- NumPy
- Numba
- SciPy
- Pandas
- Matplotlib
- mpld3
- Plotly
- transformers.Bert*
- ydata-profiling
- nltk
- spacy
- wordcloud
- ndarray (Rust)
Web & Mobile
- Ruby on Rails
- Django (Python)
- Pelican (Python)
- HTML5
- Websockets
- FME Flow Apps
- FME Data Express
- FME AR (Augmented Reality)
- Tableau
- Bulma
- OpenLayers
Databases
- PostgreSQL
- MS SQL Server
- MySQL
- Oracle
- SQLite
- ElasticSearch
- Informix
Integration
- JSON
- XML
- REST services
- REST clients
- SSL
- OAuth
- LDAP
- ActiveDirectory
- HL7 FHIR
- NEMA DICOM
- EDI X12
- FME Form
- FME Flow
- FME Flow Hosted
Tools
- Jupyter
- Docker
- git
- Pip/Poetry
- Ruff
- PyTest
- PyCharm
- RustRover
- RubyMine
- DataGrip
- Visual Studio
- NuGet
- Postman
- FME Form
- AutoCAD
- ArcGIS Pro
- Navicat
- ActiveBatch
Cloud
- AWS S3
- AWS EC2
- AWS RDS
- AWS SQS
- AWS SNS
- AWS IAM
- AWS VPC
- AWS Lambda
- AWS API Gateway
- AWS CloudWatch
- AWS Cert Manager
- Oracle Cloud Infrastructure
- DigitalOcean
GIS & Spatial
- PostGIS
- SpatiaLite
- AutoCAD
- FME
- LiDAR & Point Clouds
- ArcSDE
- ArcGIS Pro
- ArcGIS Enterprise
- ArcGIS Online
- GeoServer
- Voyager Search
- ElasticSearch
- Shapely
- NetTopologySuite
- GEOS
- NetworkX
Operating Systems
- Linux
- macOS
- Unix
- iOS
- Windows
Libraries (other than AI/ML)
- requests (Python)
- boto3 (Python)
- pyodbc (Python)
- psycopg2 (Python)
- sqlite3 (Python)
- heapdict (Python)
- aiosfstream (Python)
- fmeobjects (Python)
- certifi (Python)
- django (Python)
- elasticsearch (Python)
- urllib3 (Python)
- gdata (Python)
- ordered-float (Rust)
- Devise (Ruby)
- Pagy (Ruby)
- Nokogiri (Ruby)
APIs
- OpenAI
- ArcGIS
- Google Maps
- .NET
- Tableau
- Salesforce
- QuickBooks Online
- ChainXY
- SafeGraph
- FourSquare
- NameAPI
- SmartyStreets
- Toggl
Services
- GitHub
- GitLab
- JIRA/Confluence
- Slack
- Heroku
- OnceHub
- DigiCert
- Google Docs
- Microsoft 365
Links
- Open Source — https://kevinweller.name/resources
- Blog — https://kevinweller.name
- YouTube — https://www.youtube.com/@modgeosys
- FMEUC 2022 Presentation — https://www.youtube.com/watch?v=X2VSpW4oTRw&list;=PLbqCox4iNWPN6NlI_ipsPdVXf3alujiv8&index;=2
Current Employment
Brewer Digital (consulting to multiple industries) Denver, CO (mostly remote), 01/2023-Present — Senior Software and Data Architect with multiple roles, including:
- Solutions Architecture
- Team Management
- Software Engineering
- Data Engineering
- Data Science
- Development Team Support
- Research and Development (R&D)
- Sales and Marketing
Projects (2006 - Present)
Brewer Digital (software and data engineering), Denver, CO (remote) Senior Software and Data Architect
A private electrical engineering consultancy (engineering), Lakewood, CO (remote) 2023, 2024-Present
- Conducted deep Python static analysis, code review and coaching for their data scientist.
- Designed advanced automated spatial solution generation algorithm and proof-of-concept, plus a deep NPV (Net Present Value)-tempered ROI (Return On Investment) and breakeven analysis conservatively projecting triple multimillion returns within the first year of operation.
- Built working prototypes of several CAD automation tools using FME and Python.
- Designed, then worked with an AutoCAD plugin specialist to build the first Minimum Viable Product edition of a CAD automation tool in C#/.NET to save the client over $1 million in its first year of operation, thanks to an 80-fold increase in productivity; personally built robust concave hull algorithm implementation, and successfully resolved memory access violation issues caused by bugs in AutoCAD itself.
Internal Projects, 12/2023-Present
- Implemented AWS Lambda function with API Gateway and OpenAI GPT-4o model to provide a natural language query chatbot for an Electronic Health Records SQL database, designed specifically to prevent LLM hallucinations.
- Implemented AWS Lambda function with API Gateway to generate content using OpenAI API and GPT-4 Turbo model for award-winning FHIR-based patient portal app.
- Conducted deep Healthcare/Health Tech market research and performed statistical profiling and clustering analysis on the resulting datasets.
Medical Home Network (healthcare), Chicago, IL (remote) 2023
- Consulted and provided assistance regarding technology upgrades (Python, MySQL).
- Consulted on best practices to streamline their development and DevOps processes.
- Configured isolated batch process test environments.
- Designed and built automated testing and test results evaluation tooling.
MEDQOR (healthcare publishing and marketing), Overland Park, KS (remote) 2023-2024
- Diagnosed and resolved database performance issue of recurring batch script, resulting in 8.5 times faster processing; recommended additional architectural and performance optimizations.
- Designed complex RAG-based content generation LLM architecture.
- Fixed and documented recurring DNS and SSL certificate issues.
- Merged and cleaned up duplicate codebases.
A private human resources services company (finance) Shrewsbury, NJ (remote), 2023-2024 — Analyzed and developed automated integration architecture for periodic business-to-business ingestion of retirement account transaction and participant data into Relius database.
A private trucking IT services company (trucking and logistics) Lancaster, PA (remote), 04/2023-05/2023
- Solved an application log rotation issue in less than 3 hours that had been plaguing them for over 6 months.
- Developed modernized web/cloud-based architecture for legacy COBOL application including these advanced
features:
- Move from dying legacy Unix OS and RM/COBOL compiler to modern Linux OS and GnuCobol compiler (modified and recompiled initial program in an hour after client couldn’t do so in weeks);
- Incremental replacement of COBOL modules with Python, Ruby, and JavaScript components;
- Containerized development and deployment using shareable Docker configuration;
- Scalable cloud-based infrastructure rather than custom on-premises deployments for each customer;
- Semi-automated source code conversion;
- Replacement of all expensive commercial PDF generation and ISAM indexed file components with open source equivalents;
- Web interface to legacy text-based Curses UI with a path to piecemeal replacement of screens with modern web-based UI equivalents;
- Enablement of full integration with modern 3rd-party applications and APIs.
A private commercial real estate company (commercial real estate) Denver, CO (remote), 2023 — Developed successful web application architecture including these advanced features:
- Real-time multi-user rich text editor updates using Rails ActionCable websockets infrastructure based on an implementation of the Command design pattern, synchronizing change and comment tracking;
- Secure (OAuth and HMAC) bi-directional integration with DocuSign API.
- Assisted with the development of a practical LLM model deployment architecture.
Lipsitt Group (private clubs/hospitality services) Parker, CO (mostly remote), 07/2006-Present — Developed and currently upgrading a multi-tenant web-based reservation application and integrations for private clubs using Ruby on Rails, JavaScript, and a PostgreSQL database, all hosted on Amazon Web Services (AWS). Migrated legacy application to KVM/Qemu VM and deployed to Oracle Cloud Infrastructure when AWS environment EOL (End-Of-Life) reached.
- The application supports an administrative user interface for club personnel, including full control over their policies, reservations, and events, as well as a dashboard and extensive reporting functions.
- The application also exposes a separate user interface for club member self-service, including an event and reservations calendar, and the ability to make and manage their own reservations.
Whiteout Solutions (LiDAR geospatial services) St Johnsbury, VT (remote), 11/2022 — Built FME solution to convert LiDAR data of overgrown foliage in rights-of-way to FME AR (Augmented Reality) model for use in the field.
- Core FME workspace implements custom color coding based on point classification, as well as optional features for point cloud simplification and optimum 3D representation.
- Calling workspace executes core workspace multiple times for all required datasets and options.
CHR Solutions (Telecommunications Services) Houston, TX (remote), 07/2021-09-2021, 10/2022-Present – Built an FME and Python solution to automatically update SQL object links in whole directories of AutoCAD drawing files to reflect new database key values. CSV replacement and exception reports generated.
- Authored FME workspace with embedded Python to process a single AutoCAD DWG file into another with SQL object links updated according to a CSV conversion mapping template, and writing two CSV files, one reporting successful conversions, and the other reporting exceptions.
- Authored another FME workspace that reads a named directory and runs the first workspace for each AutoCAD DWG file therein.
- Tools used include FME Form and AutoCAD.
Mantis Innovation (commercial building and roof maintenance) Greenwood Village, CO (mostly remote), 05/2019-03/2022 – Implemented automated GIS/AutoCAD conversion system using FME, FME Flow, FME Flow Hosted, and AWS.
- Authored FME workspace to convert survey shapefiles into AutoCAD drawings in accordance with an established AutoCAD template. Drawings are rotated and scaled for best effect in page space.
- Authored another FME workspace to convert produced AutoCAD drawings into PDFs.
- Authored another FME workspace to convert produced AutoCAD drawings into PNG images.
- Authored another FME workspace to receive Amazon SQS queue messages representing uploads of new shapefiles to be converted into DWGs, PDFs, and PNGs.
- Configured an AWS S3 bucket as a shapefile inbox automatically feeding shapefile upload events to the SQS queue to kick off the conversion process.
- Configured a couple of S3 buckets to receive the converted dataset files in a standardized catalog hierarchy.
- Configured FME Flow Automations to download and process shapefiles as they are received from SQS.
- All workspaces and Automations deployed to an FME Flow Hosted environment.
SM Energy (oil & gas operator) Denver, CO (mostly remote), 01/2014-10/2015 – Built GIS self-serve data mart solution using FME, Python, and SQL Server with ArcSDE®.
- Diagnosed and fixed errors in their methane leak detection data pipeline's photogrammetry image processing logic.
- Authored two workspaces to support a clip-zip-ship solution for internal corporate use by SM Energy personnel to provide easy universal self-service access to their extensive yet disparate geodata.
- The FME Flow back-end processor accepts any combination of state and county parameters and/or an arbitrary geometry from a Geocortex front-end via custom API to specify the area of interest.
- Results are generated in the background and returned via email.
- The source data is stored in a Microsoft SQL Server database with the ESRI ArcSDE spatial extension.
- Batch jobs are scheduled and managed using ActiveBatch.
Equifax (financial information services) Denver, CO (mostly remote), 07/2018-10/2021 – Lead a team developing feature and security enhancements on a comprehensive identity theft protection services system (which I and my team originally built for an Equifax-acquired company) using Ruby on Rails, React, Docker, Microservices, custom APIs, and several third-party APIs. Applications and processes included:
- A public-facing signup application
- A custom private Customer Relationship Management (CRM) application
- A back-end batch process to detect potential identity theft risks for subscribers
- A subscriber-access portal application
- A shared library of components implementing functions common to all applications in the system
KBC Advisors (commercial real estate) Seattle, WA (remote), 08/2020-07/2021 – Built Parcel Data System using FME Flow Hosted, Python and PostGIS. ESRI ArcGIS Enterprise and ArcGIS Online consulting and high-availability deployment. Built FME/Python-based enterprise integration hub starting with AWS and Salesforce.
- Authored several FME workspaces to migrate Salesforce data to a PostGIS geodatabase, all orchestrated by an FME Flow Automation.
- Authored two FME workspaces to process Salesforce and PostgreSQL data update events on FME Flow streams.
- Developed a PostGIS geodatabase structure management subsystem with an auto-generated Salesforce data dictionary.
- Deployed an Active-Active ArcGIS Enterprise installation on AWS.
- Configured an integrated ArcGIS Online service.
Howard County Abstract & Title (title insurance) Big Spring, TX (remote), 04/2020-05/2020 — Designed and implemented solution to:
- Extract data from obsolete land abstract system into PostGIS using FME.
- Present a searchable web interface to the same data using Python and a Django web application.
CBRE (commercial real estate) Raleigh-Durham, NC (remote), 05/2019-01/2020 – Geospatial Enterprise Architect consultant with various duties.
- Integrated and automated several enterprise GIS processes and data from third party APIs using FME and Python.
- Cleansed/restructured/deduplicated data using FME and Python.
- Installed and configured the Voyager Search™ engine atop corporate geographic datasets.
- Used several tools over the course of the 7 month contract, including SQL Server with ArcSDE®, ArcGIS® Online, Portal for ArcGIS® Enterprise, ArcGIS® Interoperability Extension (essentially FME® under the covers), FME® Desktop, FME® Server, FourSquare, ChainXY, SafeGraph, and a lot of custom Python logic.
Clear Creek Royalty (mineral rights portfolio) Centennial, CO (mostly remote), 05/2019-09/2019 – Reusable extraction of mineral leasing and ownership data from third-party APIs using FME and Python.
- Authored a single FME workspace–with some embedded Python–to fulfill all functions.
Patriot Reserves (mineral rights portfolio) Denver, CO (mostly remote), 04/2018-08/2019 – Reusable extraction of mineral leasing and ownership data from third-party APIs using FME and Python.
- Authored several FME workspaces to process land records and identify individual lessees and lessors with Python fuzzy matching logic and third-party APIs.
- All workspaces deployed to FME Flow Hosted.
GeoLens, LLC (residential real estate) Broomfield, CO (mostly remote), 01/2019-05/2019 – Reusable CAD-to-GIS conversion tool implementation and enhancement using FME.
- Authored a template FME workspace as a starting point for processing custom vendor data.
- Authored FME workspaces to process data from a couple of specific vendors.
- Input drawings were error-checked, cleansed, and converted into GeoJSON with lot number labels associated with lot geometries, lot areas calculated and labeled, and exception drawings generated for incomplete, invalid (and not automatically correctable), or reprocessable geometries.
WhiteStar Corporation (geospatial data provider) Lakewood, CO (partially remote), 06/2006-05/2018 – Designed & built several self-service geospatial solutions using Python, Ruby on Rails, JavaScript, C, FME Flow, PostGIS, and SQL Server with ArcSDE®, both directly and as a team manager and architect.
- The WhiteStar Legal Mapper is a subscription-based self-serve USGS Public Land Survey System grid slicer/dicer
that interprets legal land descriptions for points (wells) or polygons (leases) and returns accurate geometries
for same. Multiple nested half- and
quarter-calls, lots and tracts are supported for each legal land description. Input data can be provided via a
custom UI or API as GeoJSON or CSV. Output is generated in GeoJSON, CSV, or Shapefile format as requested. The
solution implements a microservices architecture including:
- a PostGIS database under Amazon Web Services’ Relational Database Service;
- a Ruby on Rails web application, deployed with VueJS via a Docker container on Heroku, that exposes both a web user interface and a RESTful API;
- a Resque/Redis service to queue, and schedule batch job processing;
- a Sinatra (Ruby) legal land description processor mid-tier deployed via a Docker container by Amazon Web Services Elastic Beanstalk, using the GEOS library to aid in spatial computations and shapelib for shapefile production;
- an FME custom transformer on FME Hub for customer use;
- and an ArcGIS plugin with corresponding feature loader in Python.
- WhiteStar ClipZipShip is an extensive web-hosted self-service system to provide WhiteStar’s customers with
access to their subscribed and cleansed/accurate geodata in one of multiple formats to replace their former
labor-intensive CD-ROM shipping process. The
solution implements a microservices architecture including:
- a PostGIS database under Amazon Web Services’ Relational Database Service;
- an FME Flow installation with a complex chain of 20 FME workspaces, 22 linked custom transformers, and a system of publish-subscribe notifications to preprocess and package requested data;
- an extensive Python library to parse queries and generate SQL on top of a complex underlying data schema;
- a Ruby on Rails web application with Angular and OpenLayers JavaScript frameworks on the front end, deployed via a Docker container on Heroku;
- and an ElasticSearch full-text-search NoSQL database under Compose.io to support coordinate system selection auto-completion within the web application.
- The web application allows selection of product, order vs. quote, output coordinate system, product-specific layers, and map-selected or form-entry specification of areas of interest, which are then submitted to the FME Flow server for subsequent generation and delivery of results via subscriber email.
Identity Rehab Corporation (identity theft protection startup) Denver, CO (partially remote), 07/2007-12/2010 – Designed and managed team to build identity theft protection and management system of applications using Ruby on Rails and PostgreSQL. Applications and processes included:
- A public-facing signup application;
- A custom private Customer Relationship Management (CRM) application;
- A back-end batch process to detect potential identity theft risks for subscribers;
- A subscriber-access portal application;
- A shared library of components implementing functions common to all applications in the system.
Previous Employment
XRAD (digital radiography processor) Colorado Springs, CO (partially remote), 02/2003-07/2006 – CTO — NEMA DICOM Internet Process Architecture and Engineering.
BEA (Enterprise Integration Vendor) Denver, CO, 05/1999-06/2002 – J2EE Program Architect, Solutions Architect, Research & Development.
NewTHINK (Software Consultancy) Denver, CO, 12/1998-05/1999 – J2EE and Oracle Software Engineer.
PC Service Source (computer parts sourcing) Dallas, TX, 07/1996-11/1998 – Java GUI, 4GL Web Application, Cold Fusion Web Application, Visual Basic Systems Analyst, Software Engineer.
Lodestar Systems/Private Networks (web applications pioneer) Dallas, TX, 04/1995-07/1996 – C, C++, Perl Web Application Software Engineer.
Marken Software (accounting software) Dallas, TX, 01/1994-05/1995 – COBOL Application Software Engineer.
Tango Transport (trucking) Hall Summit, LA, 06/1992-08/1992 – DBase III Truck Dispatching Application, RPG II (IBM System/36) Fuel Certification Application Software Engineer.
Jiffy Lube (vehicle oil services) Houston, TX, 06/1991-08/1991 – COBOL, Natural/2 (4GL) (IBM mainframe) Fleet Billing Conversion Software Engineer.
Education
Texas A&M University College Station, TX, 12/1993 – BS, Computer Science + Business Analysis.