Fractional Solutions Architect

25 years of hands-on technical work. None of the full-time overhead.

Fractional Solutions Architect for AI integrations, Azure/AWS/GCP cloud systems, M365 agent deployments, and the technical decisions that compound over time.

25yr
Technical experience
IC to CEO and back to building
3
Cloud platforms
Azure MCSE / AWS since 2008 / GCP Vertex AI
$51M
P&L led as GM
While staying close to the architecture
Live
Shipping code today
SnappyClaw platform, daily commits

What a fractional SA engagement looks like

Scoped to what you actually need. None of what you don't.

Architecture review and advisory

An outside set of eyes on your current state. What is working, what is fragile, what will bite you at scale.

AI system design

Agent architectures, RAG systems, MCP integrations, multi-agent coordination. Hands-on design, not whiteboard theory.

Cloud architecture - Azure, AWS, GCP

Hands-on with all three clouds across 25 years - MCSE certified on Azure, AWS since 2008, GCP/Vertex AI for AI-native workloads. Platform-agnostic recommendation, not vendor loyalty.

API and integration architecture

System-to-system integration, event-driven architectures, API design that does not become technical debt.

Technical due diligence

For investors, acquirers, or boards who need to know what they are actually buying or building.

Engineering team mentorship

Elevating team capabilities. Code reviews, architecture coaching, and the institutional knowledge transfer that actually sticks.

The architecture I'm running today

Most architects advise from past experience. This is what I'm actively designing, building, and shipping right now - and the decisions behind it.

Application layer
Next.js 16, React 19, TypeScript, Tailwind 4, Supabase

Handles onboarding, configuration, approvals, analytics, and admin workflows. Chosen for fast product iteration, TypeScript-first development, and a pragmatic multi-tenant SaaS foundation without overbuilding infrastructure too early.

Worker layer
Python, Pydantic, HTTPX, JWT, pytest, MCP, Microsoft Agent Framework

Agent execution runs in separate worker services - not inside the web app request cycle. Agent systems need scheduled execution, retries, workflow state, long-running tasks, and fault isolation that a standard web process should not carry.

Platform layer
Docker, centralized OAuth gateway, WebSocket transport, fleet tooling

Handles integrations, runtime provisioning, transport, and operational controls. Managed runtime model with tenant isolation, predictable deployment, health checks, spend controls, and policy enforcement at the infrastructure level.

Decisions made and why they matter

Human-in-the-loop by design

The system can research, classify, draft, score, and prepare actions automatically. But anything customer-facing or externally visible routes through an approval boundary. The real problem is not generating output - it is generating output that is trustworthy, reviewable, and operationally safe.

Centralized OAuth gateway

Instead of letting each product own its token lifecycle, all OAuth and third-party connections run through a shared gateway. Calendars, email, meetings, analytics, professional networks - one place. No duplicated auth logic, no connector drift, consistent entitlements and callback handling across the platform.

Microsoft Agent Framework for new orchestration

Standardizing new orchestration work on MAF rather than accumulating framework sprawl. Better built-in orchestration primitives, more forkable, less dependent on a single closed runtime pattern. Optimizing for long-term maintainability of multi-agent systems, not speed to first demo.

Not building a thin chat wrapper

Agent-native software means the architecture separates execution from UX at the infrastructure level. The agents have different runtime characteristics than the application - different process model, different failure modes, different scaling requirements. That separation is a first-class architectural concern, not an afterthought.

Selected architectures

Shipped and used in production. Predates the current portfolio.

Azure SQL DTU Calculator

2014-2015
C#, ASP.NET, Azure SQL, JavaScript

Sizing tool for Azure SQL Database workloads. Translated on-prem SQL Server performance benchmarks into Azure DTU recommendations. Used by Microsoft field sellers and partners during the early Azure SQL migration era.

Bot Framework Training Architecture

2016-2017
C#, Node.js, Bot Framework, LUIS

Hands-on curriculum for the Microsoft Bot Framework, built as part of the Microsoft Data Science Degree program. Working code samples covering conversational AI before "agent" was a category.

Azure Data & AI Workshop Platform

2018-2021
Azure ML, Azure Data Factory, Apache Spark, Python

End-to-end hands-on workshops covering Azure Machine Learning, ADF, and Python ML pipelines. Delivered to Microsoft partners as part of the AI Accelerate and Data Science Partner Program enablement track.

Cloud Center of Excellence Reference Patterns

2018-2021
Azure platform-wide

Cloud architecture frameworks and best practices for enterprise adoption. Reference patterns the Microsoft partner ecosystem built against - covering governance, landing zones, and adoption methodology.

What people say about the engineering work

From direct reports and colleagues who worked alongside the technical architecture and engineering leadership.

His technical depth and thought leadership in Artificial Intelligence enables him to work with the technical talent on his teams in a capacity that is highly effective and impactful. I could not recommend Justin more highly as an incredible asset to any organization.

Lauren Tran

Chief Architect @ Microsoft, Office of the CTO

Reported directly to Justin

Justin is an innovator, visionary, and highly technical leader who was an honor to work with in multiple capacities. He was one of the most influential career and technical mentors I've had in my career. Justin is an outstanding people-first leader who creates an empowering environment where everyone can do their best work.

Karen Trubisky

Advisor | Former AWS, Microsoft, VMware

Justin was senior

Justin is one of the most brilliant business minds I have ever met. He has a knack for knowing how to maximize business opportunities that come his way. In addition to his business acumen, he also has incredible knowledge of technology, software, artificial intelligence and machine learning. Having spent time as a talented software engineer, Justin knows and understands the details of how to apply technology in ways others just don't comprehend.

Tony Spencer

C# Cowboy | Software Engineer

Worked on the same team

Domain expertise

Not a generalist. Deep in the areas that matter for modern AI systems.

AI and Agent Systems

  • SnappyClaw multi-agent platform (live, paying customers)
  • MCP (Model Context Protocol) integrations
  • RAG system design and implementation
  • Multi-agent coordination and orchestration
  • LLM API integration and optimization

Cloud and Infrastructure

  • Azure (MCSE certified) - AKS, AI Foundry, M365 Copilot, AD
  • AWS - EC2, Lambda, SageMaker, Bedrock, IAM
  • GCP - Vertex AI, Cloud Run, BigQuery, Gemini API
  • Multi-region, high-availability design at global scale
  • IaC, DevOps pipelines, containerization

Software Engineering

  • 25 years hands-on: C#/.NET, Python, TypeScript/JavaScript, Java
  • Distributed systems at global data center scale
  • Enterprise application architecture from ground up
  • Full product lifecycle - from first commit to production
  • Engineering management and technical mentorship

Data and Analytics

  • AI/ML pipeline architecture
  • Advanced analytics at enterprise scale
  • SQL and NoSQL design patterns
  • Real-time data architectures
  • Data Science degree

Enterprise Architecture

  • API design and governance
  • Event-driven and microservices patterns
  • Legacy modernization and platform migration
  • Technical due diligence for investors and acquirers
  • Architecture roadmaps across .NET, Java, AWS, and Azure

Product Strategy

  • Product vision and roadmap for AI-native products
  • Build vs. buy vs. partner framework decisions
  • Go-to-market strategy for technical products
  • Commercializing technology: pricing, packaging, positioning
  • Product-market fit iteration with real customers

Career evidence

From writing code to running P&Ls - and back to writing code. The full arc.

Founder & CEO|GetLatest AI / Fenix Venture
Jun 2024 - Present
  • Architecting and shipping SnappyClaw: production AI agent platform with multi-agent coordination, MCP integrations, and enterprise customer deployments.
  • Daily coding in TypeScript/Next.js. Hands-on system design for every feature.
  • GetLatest AI consulting: technical architecture reviews and AI system design for enterprise clients.
Sr. Director, Partner Ecosystem|Microsoft
Jul 2017 - Oct 2021
  • Built the technical partner ecosystem for AI/ML and advanced analytics.
  • Technical architecture advisory for the largest Microsoft AI/ML partners.
  • $100M+ in customer growth driven through architectural alignment.
Principal Software Engineer|Microsoft
Sept 2013 - Jun 2017
  • Azure adoption architecture and technical consultation for enterprise Azure migrations.
  • Led architecture review boards for strategic Data & AI customer engagements.
  • Built the Azure SQL DTU Calculator - sizing tool used by Microsoft field and partners in the early Azure SQL migration era.
  • Architected Cloud Center of Excellence reference patterns adopted across the partner ecosystem.
  • Hands-on technical depth across Azure SQL, Azure ML, Azure Data Factory, Synapse, Databricks, Cognitive Services.
Software Engineering Manager|Microsoft
Mar 2012 - Sept 2013
  • Led engineering team building massively distributed software systems for global data center infrastructure.
  • Managed architecture decisions at hardware and software boundary.
Solutions Architect|LDS Church
Jul 2008 - Mar 2012
  • Architecture roadmaps spanning .Net, Java, AWS, and Azure.
  • Multi-million dollar project delivery across enterprise systems.
  • Introduced cloud architecture patterns to the organization early in the AWS era.
Development Manager|TaxWorks
Jun 2004 - Jul 2008
  • Led development team for tax software products.
  • Built the technical foundation that supported the platform for years after.

Technical stack snapshot

Languages

  • TypeScript / JavaScript
  • C# / .NET
  • Python
  • SQL

Cloud

  • Azure (MCSE certified)
  • AWS (hands-on since 2008)
  • GCP / Vertex AI
  • Vercel / edge / Docker

AI/ML

  • OpenAI / Anthropic APIs
  • RAG architectures
  • MCP protocol
  • Agent frameworks

Data

  • PostgreSQL / Supabase
  • NoSQL patterns
  • Analytics pipelines
  • Real-time systems

How engagements work

Project-based monthly retainers. Scoped to your technical situation.

01

Architecture conversation

30 minutes to understand your technical challenge, current stack, and where things are fragile. If there's a fit, we scope it together.

02

Scoped project

Defined deliverables, timeline, and monthly retainer. Architecture review, AI system design, or ongoing advisory - scoped to what you actually need.

03

Monthly retainer

Ongoing engagement billed monthly. Adjust scope as the project evolves. No long-term lock-in - wind down when the architecture work is done.

Solve the architecture problem

30 minutes to understand your technical challenge. No pitch. No scope creep. Just honest architecture thinking.

Book a strategy conversation