Matthew Maus

I build the AI agents and automations that take manual work off HR teams: payroll, onboarding, employee questions and forms.

See the work

1,000 people. One small HR team.

Every question, change, paycheck and new hire runs through them.

Take one new hire. Monday, 9:00.

Their first day is decided weeks before they walk in.

A dozen handoffs sit behind one first day.

I automated them end to end at VitalEdge Technologies, down to booking the intro meetings. About three hours back for every new hire.

Skip the zoom

Case 01, onboarding at VitalEdge Technologies

That first day, as a system.

The zoom ended on one new hire’s first morning. This is the flow behind it at VitalEdge Technologies. Flip it to see what changed, and hover any step to inspect it.

More cases

onboarding.flow

Manual steps: 6
Time back per hire: none

Hover any step to see the tool, the guardrail and what it changes.

Same idea, five more systems.

Payroll cycle automation

Payroll, VitalEdge Technologies

Time back per cycle5+ hours

The problem

Every pay cycle, adjustments such as extra pay, vacation pay, overtime and severance were collected and keyed into ADP by hand.

What I built

An automation that gathers each cycle’s adjustments, validates them and posts them to ADP through its APIs, with HR reviewing exceptions before payroll is submitted.

Result

More than five hours back for HR Operations every payroll cycle.

HR help desk assistant

Employee support, VitalEdge Technologies

Routine questions go toThe assistant

The problem

Routine questions about policies and benefits landed in HR inboxes and pulled HR Business Partners away from higher-value work.

What I built

A chatbot that sorts each question by level, answers the ones it can, and escalates the rest to a person in HR.

Result

Many hours back for HR Business Partners, and faster answers for employees.

HR forms generator

HR operations, VitalEdge Technologies and KLDiscovery

Form creationGenerated

The problem

New hire, termination and employee change forms, such as promotions and manager changes, were created by hand in high volume.

What I built

Automated form creation on SharePoint and Power Automate, so each form is generated from existing data instead of being typed again.

Result

Many hours back for HR Operations, and more consistent forms.

Engagement comment analyzer

Employee listening, KLDiscovery

Comment codingAutomated

The problem

Open-text comments from engagement and pulse surveys for about 1,700 employees had to be read and coded by hand before leaders saw any themes.

What I built

An NLP pipeline that groups every comment into themes with representative quotes, ready for an analyst to check and present.

Result

Every comment read. Part of a listening program that contributed to a 15% rise in engagement in one year.

Calibration analytics

Performance, KLDiscovery

Regions on one dataset3 of 3

The problem

Managers in EMEA, North America and APAC calibrated performance with different data and different habits, so ratings were hard to compare.

What I built

One global dataset, calibration dashboards and guidance decks, so every manager walks into the session with the same facts.

Result

One consistent, equitable calibration process for about 1,700 employees across three regions.

Diagrams are simplified to the main steps.

How I build

An orchestrator and a team of agents: In Claude Code, one orchestrator session plans the build and directs several agents that write and test each piece. I scope the work, review what they produce and own the result.
On the systems HR already runs: ADP, SharePoint and Power BI, connected with Python, APIs and workflow tools such as Power Automate and n8n. Nothing new for employees to learn, and nothing HR has to maintain by hand.
Built solo, reviewed with IT and InfoSec: I design and build each automation myself, and partner with IT and Information Security on data governance and security compliance before anything goes live.

Claude CodePythonSQLREST APIsADPPower Automaten8nZapierMCPSharePointPower BI

Guardrails I build in from day one

A person signs off on anything that affects a person

Pay, employment changes and anything sensitive route through a human checkpoint, by design.

Reviewed with IT, InfoSec and legal

Every build goes through data governance, security and compliance review before it goes live.

Personal data stays put

Employee data stays inside approved company systems, with access limited to what each step needs.

Unsure means escalate

When the assistant cannot answer with confidence, it hands the question to a person instead of guessing.

Where I’ve done it

hours back every payroll cycle: 5+
hours back for every new hire: 3+
six-month retention after an onboarding redesign at KLDiscovery: +15%
employees supported across four countries: 1,000

People Operations Specialist

VitalEdge Technologies, remote, September 2026 to now

Automated the semi-monthly payroll cycle with Python and ADP APIs, including adjustments such as overtime, vacation pay and severance.
Automated onboarding end to end, including scheduling new hire intro meetings.
Built an AI help desk chatbot that answers routine HR questions and escalates the rest.
Built an in-house HR ticketing tool with Claude Code.
Runs payroll, benefits and leave for about 1,000 employees across the US, Canada, India and Australia.

Claude CodePythonADP APIsn8nMCPSQL

People & Culture Project Specialist

KLDiscovery, June 2023 to September 2026

Automated high-volume HR forms on SharePoint and Power Automate.
Ran engagement and pulse surveys end to end for about 1,700 employees, including NLP analysis of open-text comments.
Led performance calibration analytics across EMEA, North America and APAC.
Improved global onboarding from employee feedback, contributing to a 15% lift in six-month retention.

Power AutomateSharePointPower BINLPQualtrics

Lab Coordinator, Animal Behavior & Cognition Lab

University of Puget Sound, 2023 to 2024

Co-authored two peer-reviewed papers in the Journal of Comparative Psychology.
Ran the statistical analysis and tightened study protocols and data quality.

Research designStatisticsRSPSS

Service, hospitality and direct care

2015 to 2023

Eight years of front-line, people-facing work.
Where the employee-experience instinct comes from.

Empathy at scale

What I’d build next

Ideas, not shipped work. These are the HR problems I would point agents at next.

Offboarding orchestrationOne termination record triggers access removal, final pay, equipment return and the exit survey.
Live answers over MCPConnect the help desk assistant to HR systems so it can answer "how much PTO do I have?" from real data, with permissions.
Payroll anomaly watchFlag unusual adjustments and missing entries before each pay cycle closes.
People data quality pipelineNightly checks that catch missing or conflicting HR records before they reach a dashboard.
Learning nudgesSuggest training from role changes and skill gaps, delivered through the learning platform.
HR process audit agentMap a process from its tickets and emails, then rank each step by hours of manual work.

Start with one process.

mattmaus17@gmail.com

Résumé PDF LinkedIn

Matthew Maus

1,000 people. One small HR team.

Take one new hire. Monday, 9:00.

A dozen handoffs sit behind one first day.

That first day, as a system.

Same idea, five more systems.

Payroll cycle automation

The problem

What I built

Result

HR help desk assistant

The problem

What I built

Result

HR forms generator

The problem

What I built

Result

Engagement comment analyzer

The problem

What I built

Result

Calibration analytics

The problem

What I built

Result

How I build

Guardrails I build in from day one

A person signs off on anything that affects a person

Reviewed with IT, InfoSec and legal

Personal data stays put

Unsure means escalate

Where I’ve done it

People Operations Specialist

People & Culture Project Specialist

Lab Coordinator, Animal Behavior & Cognition Lab

Service, hospitality and direct care

What I’d build next

Start with one process.

Onboarding automation