I am a full stack engineer, applied AI builder, technical founder, and licensed Life & Health Insurance Producer across five states.

My work sits at the intersection of:

• Applied AI
• Agentic systems
• Full stack engineering
• Business automation
• FinTech
• InsurTech
• Cloud infrastructure
• Data systems
• E-commerce
• Operational software

I build beyond the frontend.

I work across architecture, databases, APIs, business logic, integrations, automation, AI orchestration, deployment, monitoring, testing, failure handling, and the operational workflows surrounding the software.

I currently run Full Stack Services LLC and build technology across multiple operating companies, including a shared AI operations platform and Learning Lantern, an adaptive AI learning and competency platform.

Next.js | TypeScript | Supabase | PostgreSQL | AI Orchestration | Microsoft Graph | Resend | Firecrawl | Google Places | OpenStreetMap | Vercel

A production CRM and automation platform built to operate real business workflows across multiple companies.

The system currently supports two independent automation pipelines.

DISCOVER
   ↓
ENRICH
   ↓
RESEARCH
   ↓
SCORE
   ↓
OUTREACH
   ↓
FOLLOW UP
   ↓
REPLY CLASSIFICATION
   ↓
BOOKING
   ↓
HUMAN CLOSE

DISCOVER
   ↓
ENRICH
   ↓
QUALIFY
   ↓
OUTREACH
   ↓
REPLY
   ↓
HUMAN REVIEW

• Automated lead discovery
• Business data enrichment
• Website crawling and extraction
• Internet research
• Identity validation
• AI lead scoring
• Evidence-backed personalization
• Automated outreach sequences
• Follow-up orchestration
• Transactional email infrastructure
• Outlook reply ingestion
• Reply classification
• Automated booking workflows
• Human escalation
• Audit trails
• Suppression controls
• Rate limiting
• Cost controls
• Automation monitoring
• Failure recovery
• Compliance protections

The two business pipelines share infrastructure where appropriate while maintaining separate scoring logic, records, timing, qualification rules, reply policies, and automation permissions.

Shared infrastructure does not mean shared risk policy.

The platform uses multiple AI providers instead of depending on one model or vendor.

Current routing includes:

Ollama
Gemini
Groq
Anthropic

Different model chains handle different tasks:

Classification
Lead Scoring
Structured Extraction
Research Interpretation
Message Generation
Data Cleanup

Provider chains can fail over based on availability and task requirements.

AI failures fail closed instead of allowing uncertain output to silently propagate through the business workflow.

AI is probabilistic. Business systems cannot be.

The models operate inside deterministic application logic, validation, transactional database operations, audit logging, rate limits, suppression rules, monitoring, and human review boundaries.

The CRM includes production controls designed around failure rather than assuming the happy path always works.

Examples include:

• Transactional email outbox
• Atomic state changes
• Duplicate-send prevention
• Interrupted-send recovery
• Advisory locking
• Retry logic
• Provider failover
• AI failure protection
• Audit history
• Daily sending limits
• Suppression by communication channel
• Identity gates for researched data
• Evidence requirements for personalized outreach
• Monitoring for failed or stalled automation
• Dead-man alerts when systems appear healthy but no useful work is happening
• Automated contract tests covering previously observed failure conditions

The goal is not simply to automate more.

The goal is to automate without losing control of the system.

learninglantern.academy

Learning Lantern is a live professional learning platform built initially around insurance education.

The platform is designed around measurable understanding rather than passive course completion.

It tracks what a learner understands, where misconceptions appear, how they reason through scenarios, and what should happen next.

• Adaptive AI instruction
• Open-ended learner checkpoints
• Sequential server-enforced progression
• Learner profiles and context
• Scenario-based evaluation
• Misconception detection
• Certification gating
• Organization administration
• Role-based access
• Consent workflows
• Privacy workflows
• Audit records
• Post-certification guidance
• Personalized prospecting workflows
• Positioning and messaging generation
• Human review boundaries
• AI guardrails

The AI is specifically restricted from inventing:

• Carrier values
• Tax outcomes
• Guarantees
• Legal conclusions
• Unsupported financial claims

The first market is insurance, where technical education, suitability, compliance, and explanation quality matter.

I also hold a Life & Health Insurance Producer license across five states, giving me direct exposure to the workflows and problems the platform is designed around.

Worked across production e-commerce and digital infrastructure supporting:

• 37+ storefronts
• 7 states
• $1M+ in monthly transaction volume
• 70,000+ product migration
• 120,000+ customer communication audience

• React development
• Shopify
• Google Cloud Platform
• Algolia
• Dutchie integrations
• Alpine IQ
• GA4
• Customer data
• Analytics
• Marketing infrastructure
• WordPress
• Laravel
• Large-scale product migration
• Production troubleshooting
• Multi-state operational systems

Working inside a regulated multi-state company reinforced one thing:

The difficult part of production engineering is rarely building the happy path.

The real work is handling:

Data integrity
Failure states
Legacy systems
Third-party APIs
Real users
Real revenue
Operational constraints
Compliance requirements
Production deadlines

fullstackservicesllc.net

I work directly with business owners to turn operational problems into software.

Projects have included:

• Custom CRM systems
• Field service software
• Scheduling systems
• Lead generation infrastructure
• Automated customer workflows
• AI operations systems
• Internal dashboards
• E-commerce platforms
• Analytics systems
• Cloud migrations
• Marketing automation
• API integrations
• Business process automation
• AI-assisted operational workflows

I am comfortable owning the entire path from problem discovery to production.

CUSTOMER CONVERSATION
        ↓
WORKFLOW DISCOVERY
        ↓
SYSTEM DESIGN
        ↓
ARCHITECTURE
        ↓
IMPLEMENTATION
        ↓
DEPLOYMENT
        ↓
MONITORING
        ↓
ITERATION

That end-to-end ownership is the part of engineering I enjoy most.

I care about systems that survive contact with reality.

That means thinking beyond whether a feature works once.

I think about questions like:

What happens when the API fails?

What happens when the model is wrong?

What happens when two jobs execute at the same time?

What happens when an email sends but the database write fails?

What happens when an external provider times out?

What happens when scraped data cannot be verified?

What happens when automation should stop?

What happens when the AI provider goes down?

What happens when a workflow partially completes?

What does the human need to see when something breaks?

The difference between a demo and a production system is usually hidden inside those questions.

Agentic Workflows LLM Orchestration AI Agents Structured Extraction Classification AI Scoring Prompt Engineering Context Engineering Human in the Loop Guardrails Model Failover Tool Calling RAG Workflow Automation

Anthropic Gemini Groq Ollama Firecrawl Microsoft Graph Google Places OpenStreetMap Resend Svix Calendly

Production AI systems, agentic workflows, orchestration, automation, guardrails, AI-integrated applications, and human-in-the-loop systems.

Financial services workflows, insurance technology, regulated systems, operational automation, and professional education platforms.

Architecture through production across frontend, backend, databases, integrations, infrastructure, and deployment.

Working directly with customers, understanding their real workflows, and turning operational problems into production software.

Replacing disconnected software and repetitive manual processes with integrated systems and autonomous workflows.

I am especially interested in engineering teams working on:

• Applied AI
• Agentic systems
• FinTech
• InsurTech
• Enterprise automation
• Forward deployed engineering
• AI-native SaaS
• Complex operational software
• Regulated AI systems
• Financial infrastructure

I am most useful in environments where the problem is not fully defined and the engineer is expected to understand the business, design the system, and ship it.

Production AI and business automation infrastructure

https://fullstack-crm-nine.vercel.app

Adaptive AI professional learning and competency platform

https://learninglantern.academy

Custom software, AI systems, and technical consulting

https://fullstackservicesllc.net