Glenn Conrad
San Antonio, TX · 210-262-8138 · gconz19@gmail.com · LinkedIn
Army veteran and full-stack software engineer, 20+ years, mostly C#/.NET with React, Angular, and Vue up front. The last two years I've been building AI into real products (document extraction, LLM pipelines with validation and a human in the loop) alongside the usual full-stack and Azure work. I work best owning a thing end to end: the database, the API, the browser, and the cloud resources it runs on.
C# .NET / ASP.NET Core SQL Server Entity Framework REST / OData / GraphQL Azure (OpenAI, Functions, Document Intelligence, Data Explorer, SQL, AD B2C/Entra) Blazor React Angular Vue / TypeScript CI/CD document-AI / LLM pipelines
Experience
Consultant Software Engineer
Contract / Sub-Contract / Freelance · San Antonio, TX · January 2016
to Present
Independent contractor. I build full-stack products end to end for a range of clients. Recent work has leaned into AI, with .NET/Azure and Vue/Angular/React underneath. Clients include Shaw Industries, USAA, JBT Corporation (AeroTech), Home Point Financial, GovTact, Conquest Consulting, and a national conference and event-services firm working with major hotel chains (since closed).
AI document-extraction pipeline for a compliance platform
- Designed and built an AI pipeline that pulls structured fields out of dense regulatory PDFs (product identifiers, classification codes, hazard and handling statements, ingredient lists, physical properties, revision dates) instead of staff keying every document by hand. It OCRs scanned files, segments the standardized document layout, and runs schema-constrained LLM extraction; then it validates the output against the governing code lists and identifier check-digits, gates on a confidence score, and sends anything shaky to a human.
- Proved it out locally first (Python + Ollama, Qwen2.5 14B on an RTX 4090), ran a cloud-vs-local bake-off, and landed the production stack on Azure OpenAI (gpt-4.1-mini) called from the C#/.NET backend, with Azure Document Intelligence for OCR. Stood up the Azure resources myself.
- On production documents it hit auto-accept quality at about half a cent each, and caught a classification code the local 14B model had missed. I costed the whole workload before committing: roughly $0.006 a document, and a ~117k-document backfill that pencils out to about $500 to $600.
- Built the surrounding product too: a barcode-first data-capture wizard tested against real scanner hardware, a date slider that shows a site's records as they stood on a past date, recursive location-tree rollups, and a two-tier auth model (tokenized read-only links for external users, Entra / Azure AD B2C with MFA for admins, so we never hold a password).
C# / .NET · Vue · Angular / TypeScript · Azure OpenAI · Azure Document Intelligence · Entra (Azure AD B2C) · Azure DevOps · Python, Ollama for the eval lab
AI room-scene generation for a flooring retailer
- Built and shipped a feature that uses generative AI to place carpet and flooring products into realistic room scenes, so shoppers and sales staff could see a product in a room instead of judging it from a swatch.
- .NET back end with a React front end. Worked it end to end, from the generation pipeline and its API surface through the customer-facing UI.
C# / .NET · React · generative AI / image generation
Legacy SOAP-to-REST API migration
- Migrated a shipping-rate integration off a carrier's retiring SOAP web service onto its new REST APIs, without dropping any option customers see at checkout.
- Reverse-engineered the legacy Perl integration (SOAP XML over LWP, XPath parsing) and wrote the migration map from key-and-password-in-body auth to OAuth2 bearer tokens, and from XML to JSON. Found that the REST rewrite splits the old single rate endpoint into two separate APIs (parcel and LTL freight) with different schemas, and confirmed the behavior against the carrier's sandbox.
- Dealt with the deployment constraints that came with it. The production server runs a 2009-era Perl, so everything had to stay compatible with it. Built a standalone test harness and planned a parallel cutover with rate-parity checks against the old system.
Perl · carrier REST APIs · OAuth2 · JSON · MySQL
IoT device monitoring + fleet-management SaaS
- Full-stack work on two SaaS products (one for IoT device monitoring, one for fleet management), with Angular front ends over .NET 8/9 and Azure Functions back ends, shipping as web, iOS, Android, and Electron desktop from one codebase.
- Folded device signal-strength and carrier data into the summary cache and moved every consumer onto a single Azure Data Explorer endpoint; standardized the alert-sending path; fixed fault-history date search.
- On the fleet side, built a delivery-mode widget and report with permission checks, then did a stretch of performance work: replaced a jQWidgets tree with Angular Material, added Angular defer blocks on the dashboard widgets, and put signal-based caching in front of repeated API calls.
Angular · TypeScript · C# · .NET 8/9 · ASP.NET Core · Azure Functions · Azure Data Explorer · Azure SQL · Azure AD B2C · GitLab CI
Earlier contract builds
- Line-of-business systems for a mix of clients, including a U.S. Navy dental-readiness system and a SaaS benefits-management platform, plus airport ground-operations software for an airport-systems company.
- Front ends in Angular/TypeScript, React, Electron, Xamarin, and WPF; back ends in ASP.NET Web API, Java/Spring, Python/Flask, Node, and Go; APIs across REST, OData, and GraphQL, with patterns from Repository up to CQRS.
- Data on SQL Server and MongoDB via Entity Framework and Dapper, with a lot of stored-proc and query tuning. Built one customer-facing store on the Umbraco API.
- Brought DevOps practice into several teams: CI on Jenkins/GitLab, TFS-to-Git migrations, deployments with Octopus/Docker/Kubernetes, and TDD + GitFlow adoption.
Earlier Experience
- GlobalSCAPE , Senior Software Engineer, 2014–2016. Led a small team on products around the core file-transfer offering; AngularJS/TypeScript UI, Web API, SQL Server, Entity Framework, custom Windows services; also ran a Java (JSF/JSP) management portal.
- iHeartMedia (Clear Channel) , Principal Software Engineer, 2014. Led an onsite and offshore team on enterprise integration (JMS/XML), billing integration, and AngularJS/WPF front ends.
- BioBridge , Software Systems Architect, 2013–2014. Stood up the team's SDLC and DevOps (Scrum, peer review, unit testing, CI, automated deploys) and built a QC tool and a defect-tracking system.
- The Scooter Store , Senior Software Developer, 2012–2013. ASP.NET, WCF, Knockout MVVM.
- Trinity Millennium Group , 2011–2012. Legacy Oracle Forms to .NET/Java modernization with custom grammar/XML tooling.
- Earlier : HCA (2009–2011) · CareNet (2007–2009) · Clear Channel (2001–2004, 2006–2007) · Texas Transplant Institute (2004–2006) · earlier roles at Immersant, Seminet (UK), and MBNA.
Education
B.A., Temple University (Philadelphia, PA)
Military Service
U.S. Army, Intelligence Analyst (1983–1988).