James E. McClurkin

IT Leader · AI Consultant · Veteran

Amount per serving

Runtime: 25+ years, zero crashes.

Serving size: one career. Ingredients: Coast Guard discipline, stagecraft precision, data-center scars, and a pathological need to understand how things actually work. The IT is invisible. The outcomes are not.

% Daily Value*

25+

Years building digital workplaces

2,500+

Users supported across 25+ year career

30+

Life sciences companies

$30k

Saved per quarter after implemented

180+

Conference rooms converted to Microsoft Teams Rooms

See the work Get in touch

Amount per serving

One operating philosophy.

Invisible IT

The system works so well no one notices it.

There is no box

The best solutions don't come from staying inside a category. They come from trusting the idea over the convention. Think it through, then build it anyway.

XLA over SLA

Measure the experience, not just the uptime.

Nutritional value

Work that shipped.

Serving size: one case study

Digital WorkplaceGlobal OpsLife Sciences

Invisible IT in Practice

Rebuilt device fulfillment and led a Zoom-to-Teams migration across 30+ life sciences companies, 1,500+ users, and four countries. Device-ready in 30 minutes. 180+ conference rooms converted. $30k saved in Q1 after implemented, and counting. The metric that matters: no one filed a ticket about it.

Amount

Value

Device-ready time

30 min

Weekly deployments

10-30

Saved per device

$400

Saved in Q1

$30k

Rooms to Teams Rooms

180+

Serving size: one case study

FrameworkXLAConsulting

Hospitality-Driven IT

Most IT organizations are Watermelon IT: green on the outside, red on the inside. The metrics hit every target while the experience underneath stays broken. HDIT (Hospitality-Driven IT) fixes that by replacing SLA (Service Level Agreement) thinking with XLA (Experience Level Agreement): measuring how an interaction actually felt, not just whether the ticket closed. Built on Marco Gianotten's Output/Outcome/Impact model, grounded in HappySignals benchmark data, and informed by Will Guidara's distinction between service and hospitality in his book Unreasonable Hospitality. The framework is documented, applied in practice, and backed by 25+ years of outcomes.

Amount

Value

SLA to XLA

Shifted

Framework status

Complete

Content series drafted

5 posts

Years of outcomes behind it

25+

Serving size: one case study

AI SystemsInfrastructure

Deep Thought

The problem: AI assistants forget everything when the session ends. Deep Thought fixes that with persistent memory, version-controlled skills, a live project tracker, on-demand archiving of chats and file content, and an idea pipeline that works across Claude, OpenAI, Gemini, and whatever comes next. Career OS, Mad Skills Library, and Content Forge all went from tracker entries to versioned, deployed systems without losing a single thread. Built the tool. Proved it by using it.

Amount

Value

Session continuity

AI platforms supported

3+

Serving size: one case study

AI SystemsSkill Engineering

Mad Skills Library

Built the version-control system that governs every Claude skill across all active projects. A Notion-backed registry with two enforcement skills at its core. skill-creator drafts and tests new skills through subagent evaluation before they ship. skill-update archives the prior version and increments the new one, automatically gated before any file is delivered. The system that builds and versions its own instructions.

Amount

Value

Distinct skills tracked

122

Tracked revisions

144

Projects with documented reuse

3

Unversioned releases

0

Serving size: one case study

AI Product DesignMulti-Persona

Claude Project Creator

Most Claude prompts are written in 20 minutes and forgotten in 40. This builds them properly. Seven expert personas interrogate the idea before a single line ships. On-demand Subject Matter Experts (SMEs) weigh in. Three intake modes, mandatory Plan Mode before any build, one ZIP delivered with four deployment-ready packages (Chat, Project, Cowork, Code). The difference between a prompt and a product. Career OS and Content Forge were both built using CPC.

Amount

Value

SME personas per build

7

Deployment targets

4

Serving size: one case study

AI EngineeringPersonal Tool

Career OS Built using CPC

Built an AI-native career operating system on Claude, then used it as my own job search engine. Notion as hub, three Claude deployment targets as spokes. The ghost-job detector caught a role that had been reposted for nearly a year. Three versions shipped in under two weeks. 85%+ ATS score on a tailored resume.

Amount

Value

Versions shipped

3

Time to ship

<2 weeks

ATS compatibility score

85%+

Architecture

Hub-and-spoke

Serving size: one case study

AI ContentSocial Systems

Content Forge Built using CPC

Not a prompt. A standing content operation. Seven specialist personas on call, platform-native adaptation for LinkedIn, Instagram, and Facebook, carousel PDFs via a Python/reportlab pipeline, campaign arc construction, a three-part hashtag audit, and a hard two-attempt ceiling before any task loops. Built to run, not to be babysat.

Amount

Value

On-call specialist personas

7

Platforms adapted

3

The recipes

Ingredients.

Feb 2022 — Apr 2026

Director, IT — End User Services and Digital Workplace Experience

Flagship Pioneering · Cambridge, MA

Owned the full end-user experience across 30+ life sciences companies, 1,500+ users, and four countries. Rebuilt device fulfillment from the ground up: 30-minute device-ready time, 10-30 weekly deployments, $30k saved in Q1 after implemented. Led the Zoom-to-Teams migration, converting 180+ conference rooms. Also led the full IT team of 8 technicians in an interim capacity for approximately 8 months, designing the enterprise MDM framework adopted as the portfolio standard.

Managed a $3.5M annual budget across hardware, services, and vendor contracts. Served as hands-on Okta Certified Administrator across the portfolio, supporting identity governance, SSO integrations, and lifecycle management for 1,500+ users. Built AI-assisted workflows across service delivery and documentation, including an employee goal-setting assistant adopted and extended by HR. That was the organization's first repeatable model for AI adoption among non-technical users.

Digital WorkplaceGlobal OpsLife SciencesOktaMDM

Jul 2019 — Feb 2022

Senior manager, IT

Inari Agriculture · Cambridge, MA

Owned IT operations for a fast-growing agricultural biotech startup through rapid headcount growth from 60 to 450 employees. Managed a $6.5M annual IT/SaaS budget and built a 9-person distributed team across Boston, Indiana, and Belgium, replacing full MSP dependency with an in-house operation without the infrastructure seams showing. The budget grew. The chaos didn't.

Designed and implemented the enterprise MDM framework and zero-touch onboarding automation from the ground up: new hires logged in with their credentials and were fully provisioned and productive the same day. Built identity and access foundations as Okta Certified Administrator across Okta Workflows, Azure AD, and UEM/MDM, automating the employee lifecycle and hardening endpoint security across a multi-site, multi-OS environment. Implemented a cross-departmental ticketing and escalation system that improved escalation efficiency by 25% and reporting accuracy by 20%. Executed COVID-19 remote work continuity for 120 employees with zero downtime during high growth.

Scale-upBudget OwnershipAgTechOktaZero-Touch

Jan 2018 — Jul 2019

Facilities and IT manager · Acting construction PM

Greentown Labs · Somerville, MA

IT and facilities for the world's largest clean energy incubator: 300+ members, 100,000 sq ft across three sites. Mid-tenure, stepped in as acting construction PM for a $12M headquarters buildout. All on time and on budget.

Implemented a Jira-based ticketing system to standardize facility and IT service delivery across a complex, multi-tenant environment. Managed building systems, vendor contracts, and IT infrastructure across all three sites simultaneously.

CleanTechFacilitiesConstruction PM

Earlier

Enlisted service member

United States Coast Guard

Boat Captain. Search and rescue. Firefighting. Drug enforcement. EMT. Lighthouse Keeper. When the people and property in your care depend on the decision you make in the next thirty seconds, you stop guessing and start knowing. Everything since has been a much more forgiving version of that.

Boat CaptainSearch & RescueEMTLighthouse Keeper

Amount per serving

Skills and expertise

% Daily Value*

AI advisory and strategy

93g

Context Engineering 35g

87%

AI Readiness and Governance 30g

92%

Human-First, Tech That Follows 28g

110%

IT leadership and digital workplace

85g

Digital Workplace Strategy 35g

95%

Global Operations at Scale 28g

88%

Before You Had to Ask 22g

110%

Operations and service delivery

68g

ITSM and Experience Delivery 28g

88%

SaaS and Procurement 25g

82%

Caught It Before It Broke 15g

110%

* Percent Daily Values are based on a 2,000g technology diet. Your professional needs may be higher or lower.

In development

What's next on the table.

Ingredients being finalized

ConceptPre-Launch

VetsProject (working name)

Veterans already know how to execute under pressure, work as a team, and build things that hold. VetsProject puts that to work with advanced manufacturing training, a business-ownership track, and a deployment model designed for moments when communities need skilled hands fast: temporary housing, service buildings, and essential infrastructure in the aftermath of tornadoes, floods, and disasters. Built accessibility-first for veterans with service-connected disabilities. The training architecture is ready. The first cohort forms when the resources do.

Ingredients being finalized

ConceptProposed Pilot

Music Heals

Music is already in the treatment room. This makes it the right music. A proposal for free streaming during active treatment: cancer, chronic illness, or any condition that brings someone back to that room, again and again. Linked to care schedules so a patient's own playlist is ready when they walk in. A companion community lets patients share what got them through a hard session, anonymously or by name. Patient data secured and HIPAA compliance built into the architecture from day one. The idea is right. The pilot is the next step.

Nutrition facts

What Good Looks Like.

25+ years building digital workplace and end-user services organizations that people never have to think about. The career spans Coast Guard operations, theatre stage management, and hands-on IT leadership across biotech, agtech, and cleantech. An unusual path that turned out to be exactly the right one.

What Good Looks Like LLC

What Good Looks Like LLC sits at the intersection of operational discipline and human experience. It covers the full spectrum: hospitality-driven service delivery, AI tools that actually work for the people using them, and the operating frameworks that connect the two. HDIT is one spoke. The AI toolkit is another. The through-line is the same across all of it: stop measuring activity and start measuring whether people feel good about what happened.

HDIT — Hospitality-Driven IT

A framework borrowed from an industry that has always measured success by how the person felt, not whether the process closed on time. Hospitality-Driven IT applies that standard to technology delivery, shifting the operating question from "was the ticket resolved?" to "did the person leave better off than they arrived?" XLA over SLA. Experience over compliance (XLA).

Background

Coast Guard. Stage manager. Lighthouse keeper. IT director. The path doesn't follow a straight line, and that's the point. Twenty-five years of building things that work quietly, across environments where the margin for error ranged from a missed cue to a missing boat. The through-line is the same: show up, figure it out, don't leave a mess.

Manufactured in Boston, MA

Amount per serving

Get in touch

Email james@jamesmcclurkin.com LinkedIn linkedin.com/in/jamesmcclurkin Send a message

* This label is provided for informational purposes. Actual response time may vary based on time of inquiry.