Principal Platform Engineer

Matthew Branche

I turn fragmented infrastructure workflows into secure, documented services that engineering teams can adopt and operate.

View selected work Download resume

969service requests in ten months

300+automation jobs supported

11sites using shared monitoring patterns

What I do

Build the service around the system

The technology matters, but adoption depends on the whole operating model: clear intake, reusable automation, secure runtime configuration, useful observability, durable documentation, and honest ownership boundaries. That is the layer I build.

Selected work

Platforms made usable

Sanitized case studies based on verified artifacts and aggregate outcomes. No employer source, topology, or confidential data is shown.

Service intake platform

From a team homepage to a shared engineering service

Created a one-stop lab services portal, added FastAPI and Jira-backed intake, and helped scale the service pattern with additional engineering groups.

Measured result969 requestsDec 2025 through Sep 2026

Core patternForms + APIConfigurable request routing

LifecycleLocal to sharedMultiple release generations

Shared problem: Requesters faced fragmented service links, knowledge, and team-specific intake paths.
My contribution: Built the substantive original portal and FastAPI intake implementation; maintained releases, integration behavior, and audience-specific documentation.
Collaborators: Partner service groups, platform contributors, product owners, and later maintainers shaped and scaled the shared result.
Collective result: A centralized interface connected human and programmatic requests to routed engineering backlogs.

Multi-site observability

Beyond reachability checks

Helped evolve basic ping monitoring into a reusable Prometheus and Grafana pattern spanning infrastructure classes and 11 lab sites.

Contribution and evidence

The core deployment repository and stack are verified and were authored by a collaborator. My contribution was deployment, expansion, asset onboarding, and operationalization. The 11-site count is an operationally verified figure.

Secure automation foundations

Secrets that automation could safely consume

Scaled a containerized Vault capability from one development team and 12-15 active users into a central private service supporting 6+ teams, approximately 50 users, and 10 service accounts.

Contribution and evidence

Both deployment generations are supported by repository history. I authored the inspected successor release and operating baseline, including enterprise authentication and CI/CD integration patterns; organizational operation remains collaborative.

Infrastructure source of truth

Planning from current state to proposed state

Customized and deployed NetBox with enterprise authentication and inventory synchronization for rack elevation and infrastructure planning.

Contribution and evidence

The customization release and deployment work are verified. The foundation includes upstream community work and initial contributions from colleagues, which remain part of the attribution.

Hybrid AI systems

Prototype freely at home. Operationalize responsibly at work.

My AI practice spans two deliberately separate environments. I use a private home lab to learn quickly, then carry the engineering patterns into approved enterprise tools without moving home data, credentials, models, or code across that boundary.

01GPU computeSelf-hosted NVIDIA acceleration

02Models and agentsLocal inference and tool use

03Atlas memoryBoot context, notes, semantic search

04Specialized servicesAutomation and market experimentation

05Enterprise transferPatterns rebuilt with work-native controls

Atlas today

A private AI operating layer

Atlas is my home platform for exploring persistent agent memory, authenticated service APIs, local model execution, and reusable tool interfaces. The current artifacts verify memory boot context, note retrieval, semantic search, and access-control testing.

Building toward

Specialists coordinated as one system

The longer-term direction is an agentic operating layer that can coordinate whatever I want to build at home, including infrastructure automation and an experimental stock and crypto research/trading service. This describes an active build direction, not a claim of investment performance.

Transfer to work

Learning moves. Boundaries stay.

I bring over architecture lessons: service boundaries, memory patterns, tool contracts, evaluations, observability, secrets handling, human approval, and failure recovery. At work, those patterns are implemented with employer-approved AI, identity, data, repositories, and deployment controls.

Engineering approach

A contribution model built for shared systems

Platform work is almost never a solo artifact. I separate the problem, implementation contribution, collaborators, collective outcome, and evidence so the story remains useful without erasing anyone else.

01Shared problem
Define the recurring friction and the people it affects.
02My contribution
Name the architecture, implementation, or operational work I performed.
03Collaborators
Preserve authorship, upstream foundations, and partner-team roles.
04Collective result
Describe the capability the organization gained.
05Evidence
Anchor the claim in artifacts, history, adoption, and measured outcomes.

Experience

Thirteen years across lab services and platforms

Official titles are shown as held; the platform-engineering headline describes the work and target direction.

2019 - Present

Dell Technologies

Lab Support Principal Engineer

Assignments across software release engineering and lab services: internal platforms, automation, service intake, secrets management, observability, infrastructure planning, and team enablement.

2017 - 2019

Brocade, a Broadcom Company

Senior Global Service Lab Engineer

Self-service lab access, automation, power and utilization workflows, customer proof-of-concept environments, and technical enablement.

2013 - 2017

Brocade

Lab Services Engineer

Lab access, automation, demonstration workflows, testing environments, and operating documentation for global technical teams.

Technical range

Software, infrastructure, and the operating layer between them

Build

Python, FastAPI, REST APIs, Bash, reusable libraries, testing

Run

Docker, Jenkins, GitHub, Jira, Artifactory, NGINX, Vault

Observe

Prometheus, Grafana, SNMP, cAdvisor, infrastructure exporters

Model

NetBox, Linux, Windows, VMware, server and network infrastructure

Accelerate

NVIDIA GPUs, local inference, persistent memory, agent orchestration

Govern

Human approval, evaluations, secrets, auditability, environment boundaries

Contact

Building the next useful system

I am exploring senior and principal platform engineering roles where hands-on building, infrastructure judgment, and durable service design all matter.

mbranche@gmail.com