What I build

Software and platform work

The application side: what gets designed, written, and shipped.

AI Integration & LLM Consulting

Add AI features that hold up in production, built across the major providers.

  • Multi-provider LLM orchestration (OpenAI, Anthropic Claude, Google Gemini, Azure OpenAI)
  • Prompt engineering with versioning and A/B testing
  • Streaming responses with real-time chat interfaces
  • Session management and persistent conversation history
  • Per-tenant usage analytics and cost optimization
  • MCP (Model Context Protocol) server implementation
  • Speech-to-text and text-to-speech integration

API-to-MCP Integration

Make your existing APIs usable by AI assistants and agents. With Kuhn Consulting's API2MCP platform, your REST APIs become Model Context Protocol (MCP) servers, so tools like Claude can call your services directly. No custom integration code to build or maintain.

  • Expose any REST API as a hosted MCP server and toolset
  • Auto-generated tools from your OpenAPI spec or existing endpoints
  • Tools stay in sync as your API changes
  • Secure per-provider credential and auth management
  • Centralized dashboard for servers, tools, and usage analytics
  • Ready for Claude, Claude Code, and other MCP clients

SaaS Platform Architecture

Launch a SaaS product that can scale: multi-tenant architecture built for growth, security, and the day-to-day work of running it.

  • Multi-tenant architecture with database-level tenant isolation
  • Subscription billing with full lifecycle management
  • Rate limiting, IP restrictions, and abuse prevention
  • Comprehensive audit logging for all critical actions
  • Tier-based access control with license token management
  • Real-time messaging via SignalR WebSockets
  • RESTful API design with OpenAPI/Swagger documentation

Full-Stack Web Development

Fast, reliable web applications, built on the proven Microsoft .NET stack.

  • ASP.NET Core Web APIs and Razor Pages applications
  • Blazor Server and Blazor Hybrid interactive UIs
  • React with TypeScript for SPA front-ends
  • SQL Server database design, optimization, and migrations
  • Entity Framework Core with complex data modeling
  • Reusable component library development
  • SSRS reporting and SSIS data pipelines

Cross-Platform App Development

Native apps for Windows, iOS, macOS, and Android without a separate codebase for each one. I build in whichever stack fits the project: .NET MAUI when the rest of the system is already C#, React Native with TypeScript when the team lives in JavaScript.

  • .NET MAUI for Windows, iOS, macOS, and Android
  • React Native with TypeScript for iOS and Android
  • Blazor Hybrid for shared web/native UI components
  • Local-first architecture with SQLite for offline support
  • Mapbox integration for location-based features
  • Platform-specific APIs and native capabilities
  • App store deployment and distribution

Payment Integration

Take payments without the headaches: Stripe, PayPal, and in-app purchases, with the full subscription lifecycle handled.

  • Stripe integration with subscription lifecycle management
  • PayPal checkout and recurring billing
  • Apple and Google In-App Purchase support
  • Create, pause, resume, cancel, and refund workflows
  • Transaction review with admin flagging
  • Detailed audit trails for all payment actions

Authentication & Security

Protect your users and their data with modern authentication: passwordless login, multi-factor security, and enterprise SSO.

  • OAuth2 with social login (Google, Microsoft, Facebook)
  • WebAuthn / Passkeys for passwordless authentication
  • Two-factor authentication with recovery codes
  • JWT session management with refresh tokens
  • SAML SSO for enterprise client integration
  • Column-level database encryption

Communication Services

Keep users in the loop with reliable, compliant messaging: email, SMS, and real-time notifications across multiple providers.

  • Email services via MailKit with HTML templates
  • SMS via Twilio, AWS SNS, and Vonage
  • Batch processing for bulk communications
  • Opt-out compliance and preference management
  • Rule-based automated notification systems
  • Real-time messaging with SignalR

E-Learning & SCORM

Build or enhance learning management platforms with 20+ years of e-learning industry expertise and standards compliance.

  • SCORM 1.2 and SCORM 2004 compliance
  • AICC courseware integration
  • Learning Management System development
  • Courseware conversion and standards consulting
  • Enterprise LMS platforms for Fortune 500 clients

Home & Small Business IT (Vail Valley)

The local, hands-on side of the practice. On-site and remote technology work for homes and small businesses across Eagle County, Colorado, from Vail down to Glenwood Springs.

  • Small business networking, Wi-Fi, and firewalls
  • Workstation setup, upgrades, and maintenance
  • NAS file storage and tested backup
  • Practical AI tools for small teams
  • Home theater, automation, and phone setup
Infrastructure and operations

Building it is only half the job

Someone still has to run the servers, ship the releases, and keep the whole thing healthy. On most engagements that's me too.

Deployment & Hosting Infrastructure

Code that only runs on a laptop isn't finished. I build the hosting underneath an application and then run it in production.

  • Linux server administration, from provisioning through day-to-day operations
  • Kubernetes and lightweight k3s clusters, with per-app namespaces, resource limits, and health probes
  • Containerized apps served behind nginx and Traefik
  • Separate staging and production environments running from one build
  • Persistent storage, scheduled jobs, and in-cluster databases
  • Moving live workloads off managed cloud services without losing data

CI/CD & Automated Testing

Releases should be boring. Every change builds, tests, and deploys the same way, so shipping stops being an event.

  • Build and test pipelines that run on every pull request
  • Unit and functional suites kept separate, with results archived per run
  • Browser-level end-to-end tests that exercise real user flows
  • Merges gated on a green pipeline, so nothing lands untested
  • Automatic deploy to staging, then a manual promotion of that exact verified build to production
  • Database migrations applied by the pipeline, so environments don't drift
  • Health-checked rollouts with one-command rollback

Network Security & Secure Access

The safest port is the one that was never opened. Access gets designed in, not bolted on after the first incident.

  • Servers with no inbound ports exposed to the internet, reached over an encrypted tunnel
  • VPN and Zero Trust remote access, so internal tools ask who you are instead of just where you came from
  • Access policies with named groups and per-application rules
  • Network policies inside the cluster, so one app can't reach another's services
  • TLS, HSTS, and security headers set deliberately rather than copied from a blog post
  • Automated certificate issuance and renewal for customer domains
  • Least-privilege database logins and encrypted secrets that never land in the repo

Monitoring, Backup & Recovery

Finding out something broke before your customers do, and being able to put it back.

  • Metrics, alert rules, and uptime checks tuned to fire on real problems instead of noise
  • Nightly database backups, checked for integrity rather than assumed good
  • Restore drills against a scratch copy, so recovery is proven and not theoretical
  • Storage that retains your data even when the thing using it gets deleted
  • Health endpoints and structured logs that keep an incident short
  • Runbooks covering how to deploy, how to roll back, and who can reach what

Developer Platform & AI-Assisted Delivery

Most of the time lost on a project goes to integration and onboarding. This is the part that goes after both.

  • A git host and issue tracker, self-hosted or wired into what your team already uses
  • Issues connected to the coding workflow, so work gets claimed, planned, and moved through review in one place
  • AI harnesses that implement a plan in stages and review each stage before moving on, instead of dropping one enormous change on you
  • AI skills that encode your API, your conventions, and your deploy steps, so an assistant works the way your team already does
  • Your internal API exposed as AI-native tools, so assistants call your systems instead of guessing at them
  • Repo structure and documentation that gets a new engineer oriented quickly
Common questions

Frequently asked questions

Straight answers about how engagements work.

What does AI integration consulting involve?

AI integration consulting means connecting your application to large language model (LLM) APIs (OpenAI, Anthropic Claude, Google Gemini, or Azure OpenAI) and building the infrastructure around them: prompt engineering, streaming responses, session and conversation history, cost controls, and per-user usage analytics. Kuhn Consulting builds production-grade implementations, not prototypes.

What is Model Context Protocol (MCP) and why does it matter?

Model Context Protocol (MCP) is an open standard that lets AI assistants like Claude call external tools and APIs directly. Instead of users copy-pasting data between their tools and an AI, MCP lets the AI query your systems on demand. Kuhn Consulting builds MCP servers that expose your REST APIs as AI-native tools, and also built API2MCP, a platform that automates this for any OpenAPI-documented service.

How long does it take to build a production SaaS backend?

A minimum viable SaaS backend with authentication, multi-tenancy, subscription billing, and a basic API typically takes 8 to 16 weeks depending on complexity. Kuhn Consulting speeds that up with the Wildwood Platform, a production-tested SaaS backend that already has these pieces built, so new products launch on enterprise-grade infrastructure in a fraction of the time.

Do you handle hosting and deployment, or just write the code?

Both. Kuhn Consulting builds the application and the infrastructure it runs on: Linux servers, container hosting on Kubernetes, separate staging and production environments, TLS and access control, backups, and monitoring. If you already have a platform team, I work inside what you have. If you don't, I set it up and hand you the runbook that explains it.

What does a CI/CD pipeline actually do for a small team?

It takes the manual steps out of releasing, which is where releases usually go wrong. Every pull request builds and runs the test suite, and a merge only happens if that passes. The change then deploys to staging on its own, and the same verified build gets promoted to production when you're ready. Database migrations run as part of the deploy so environments don't drift, and a bad release rolls back with one command.

How do you get a team set up for AI-assisted development?

Most of it comes down to context, because an AI assistant is only as useful as what it knows about your systems. Kuhn Consulting wires your issue tracker and repository into the development workflow, encodes your API and your conventions as reusable AI skills, and exposes internal services as AI-native tools the assistant can call. The same groundwork that makes an assistant productive makes a new engineer productive, so onboarding tends to get easier as a side effect.

Do you work remotely?

Yes. Kuhn Consulting is based in Eagle, Colorado and works remotely with clients across the United States. Every engagement is handled directly by Mark Kuhn, with no subcontractors or junior developers.

Get in touch

Ready to start a project?

Tell me what you're building, and let's figure out if it's a fit.