I lead complex software programs and stay close enough to the technology to understand what I am asking teams to build.

12+ years across enterprise SaaS, cloud, infrastructure, IoT, integrations, and technical program management. Most recently, I led work across 12 Scrum teams on a .NET MAUI platform at Neptune Technology Group, coordinating engineering, QA, product, CloudOps, dependencies, delivery risk, and executive reporting.

Outside the day job, I build and operate production software through Imogen Labs. That work gives me firsthand experience with architecture, APIs, containers, authentication, CI/CD, observability, security, infrastructure, AI-assisted development, and the tradeoffs engineering teams make every day.

Target roles: Senior / Principal Technical Program Manager Focus: AI · Enterprise SaaS · Cloud · Infrastructure · Developer Platforms

mjashley.com · LinkedIn · michael@mjashley.com

I have spent more than a decade working where engineering execution, product priorities, and business outcomes have to line up.

At Neptune Technology Group, I ran program execution across 12 Scrum teams delivering software for connected-device and cloud platforms. Before that, I led SaaS programs at Healthcare Integrations and technical projects at G5 TEK.

The work is familiar: roadmap alignment, sprint execution, cross-team dependencies, risk management, release planning, stakeholder communication, executive reporting, and enough technical depth to know when a problem is a schedule issue, an architecture issue, or both.

PSM · CSPO · SA

I do not build software because I am trying to leave program management.

I build because it makes me better at it.

Through Imogen Labs, I have taken products from idea to production across infrastructure, AI, mobile, storage, backup, developer tooling, and open source. I write requirements, define architecture, manage backlogs, use AI-assisted implementation, review the running product, test failure cases, run security reviews, ship releases, and operate the systems afterward.

That means when an engineering team talks about container networking, authentication, deployment pipelines, observability, API contracts, technical debt, model behavior, or production failure modes, I do not need the simplified version.

I use AI-assisted engineering extensively to increase implementation velocity. Product definition, architecture, sequencing, acceptance criteria, validation, security, release quality, and production ownership remain mine.

Every meaningful change starts with written work, moves through a real delivery pipeline, is reviewed against a running environment, and is expected to survive production rather than merely pass a demo.

A self-hosted infrastructure platform built around one Go binary. Docker, KVM/QEMU virtual machines, LXC, storage, SMB/NFS shares, infrastructure management, and local AI tooling.

My own production infrastructure runs on Eight.ly OS after replacing the Proxmox and Unraid environments it grew out of.

Deduplicating backup with content-defined chunking, AES-256-GCM encryption, retention policies, and integrity verification.

Remote Docker management for systems you are not sitting in front of. Apache 2.0, signed binaries, provenance, SBOMs, and multiple security reviews.

A self-hosted homelab dashboard with 117 applications, multiple deployment modes, a live demo, and an iOS companion app.

A native SwiftUI wellness app with AI chat, grounding and breathing exercises, brain games, and HealthKit integration. Live on the App Store.

An AI operator scaffold with identity, layered memory, scheduled jobs, and multi-agent roles. Published on npm. The underlying methodology, IPAC, is public.

Watches Docker events and automatically creates Cloudflare DNS records for Traefik-labelled containers.

Not everything here lives in repositories I control.

• OmniVoice — Intel XPU / Intel Arc GPU inference support — upstream contribution adding Intel GPU support.
• Apple App Store — production iOS applications shipped publicly.
• Docker Hub / GHCR / npm — distributed software and packages used outside my local environment.
• Open source — production code, documentation, releases, and issues are public rather than represented only as portfolio screenshots.

An SDLC, not a story about one. Every change starts as tracked work with acceptance criteria, moves through a sprint or delivery queue, and stops for human validation before production. Architecture decisions, runbooks, and process documentation live alongside the products they support.

Architecture and accountability stay human-owned. AI can implement quickly. It does not get to decide what “done” means. Scope, sequencing, architecture, acceptance criteria, browser review, security review, and production ownership remain human responsibilities.

Guardrails have to prove they work. Automated gates block bad releases for formatting, oversized files, undocumented work, failed tests, or unverified claims. A guardrail is only useful if it has been tested against a real failure case.

I operate what I ship. My infrastructure runs on the software I build: bare-metal nodes, Docker, reverse proxying, Cloudflare tunnels, self-hosted GitHub Actions runners, multi-architecture builds, monitoring, storage, and backup.

Open to Senior and Principal Technical Program Manager roles in AI, enterprise software, cloud infrastructure, and developer platforms.

mjashley.com · LinkedIn · Email

_{Counts above are refreshed daily from their sources by a workflow in this repo. Last checked 2026-09-20.}