Product Management Executive · Enterprise SaaS
I build products. Now I build them with AI.
I take complex products to market. Strategy, roadmaps, and launches shipped across multiple industries in both B2B and B2C SaaS, including 100+ enterprise integrations where the data is regulated and nothing is allowed to break. I never stopped building, and today I direct AI coding agents to take a product from idea to a real release.
Open to Senior Product Roles and AI advisory for companies building or scaling with AI.
Burlington, ON
I began my career writing code, moved into business analysis, and grew into senior product leadership. Along the way I owned product strategy and a $2M budget at Limelight, leading the core SaaS platform rebuild from concept to launch, ran a portfolio of 14+ product lines contributing $5M+ ARR at Points.com, and built and led product and integration teams of up to 10 at Limelight, IMS, ADP Canada and MealSuite. That work spans healthcare, fintech, insurance, edtech, cybersecurity, martech, HR tech and grocery retail.
The developer never left. I still get into the details, sketch the data model, and build the prototype myself. That is why AI clicked for me. I use it to research, design and build products end to end, and I bring the same discipline to how teams scope, evaluate and release.
AI is also my day-to-day product operating system. I use it to synthesize research, draft and iterate PRDs, pressure-test roadmaps, and build prototypes before engineering sees a ticket. When I lead teams, I apply the same AI product discipline: scope what actually needs an LLM, evaluate model and prompt output against real cases, route the right model to the right task, and keep a human in the loop anywhere a bad output has real cost.
Strategy, roadmaps, portfolio and P&L ownership, and scaling products from zero to one across regulated B2B and B2C SaaS.
HL7 and EHR, APIs, and platform integrations at scale.
Directing AI coding agents to take products from idea to working build solo. Agentic workflows, RAG, model evaluation and human-in-the-loop design.
Commercialization, product-led growth, and partner ecosystems.
Experience
Integrating senior-living clinical and partner feeds is complex, costly, and slow to deliver.
100+ Integrations · 65+ Days to Under 30Brands needed one platform to run experiential marketing programs end-to-end with full CRM integration.
100+ Global Brands · Months to DaysData throughput limited how fast and how deep sports content could be published.
+150% Data Speed and VolumeAuto insurers wanted claims to start the moment a crash happened, not days later.
Industry First · Team of 5 PMsCurriculum content updates took months to reach millions of learners.
50M+ Students · Months to HoursPartners had no supported way to build on the platform.
5 Partners Certified · 8 Integrations LiveCanada's independent life insurance channel runs on back office platforms that are hard to modernize.
National Back Office PlatformA security services business needed productized offerings to grow.
200% YoY Growth · 2 to 15 OfferingsA large HR platform needed a roadmap that moved thousands of customers forward.
2,000+ Customers · 3 ReleasesThree product lines needed to ship faster without losing budget discipline.
200%+ Development Efficiency · P&L OwnershipLoyalty currency commerce needed new revenue-generating products.
$5M+ ARR · 14+ Product LinesCommunity
Products
Anyone can prompt a chatbot. These are working products. The usual path takes a staffed team, designers, front end, back end, QA, release engineering, and a run of sprints. I built these solo by directing AI coding agents, and the first working build of Fairway was in front of testers in under a business day.
This is my operating model, not a side hobby. One product leader, directing agents, taking an idea to a real release: architecture, auth, data model, observability, and an automated pipeline to the app stores. It is the same speed I bring to the teams I work with.
Product lead and AI orchestrator, no hand-written code
Groups still keep score on one paper card, then someone has to add it all up and retype it into an app afterward.
0→1 In Testers' HandsProduct lead and AI orchestrator, no hand-written code
AI job-search tools embellish. Candidates need tailoring that stays true.
0→1 In Testers' HandsProduct lead and AI orchestrator, research through build
Smaller post-acute-care SaaS vendors lose deals over the cost and time of legacy EHR feed integration.
0→1 BuiltOperating Model
Most PMs who say they use AI mean they prompt a chatbot. I design and run agentic systems. To prove it, I turned my own job search into one and ran it on myself for months.
This is dogfooding. Before I ask a team to trust an autonomous loop, I run one where a bad output costs me, not a customer. It runs the whole search end to end across real tools, with memory that compounds. Same way I build product: find where AI earns trust, keep a human in the loop where it matters, and redesign the workflow instead of bolting AI onto the side.
Searches the web and job boards for roles matching a multi-lane target, not one keyword. Normalizes every posting into one schema.
Scores each role for fit, flags stretch vs. strong matches, and filters out noise before I spend attention.
Generates a role-specific resume and cover letter for every posting from a bank of my own pre-approved answers. Auto-names and exports to PDF.
Keeps a live board as the single source of truth, with every opportunity tracked across stages, follow-ups, and next steps updated automatically.
Drafts recruiter and hiring-manager outreach and tracks headhunter follow-ups as their own pipeline.
Assembles role-specific prep, including an AI-fluency reference before technical screens.
Writes durable memory between sessions and logs finished answers back to the bank. Each cycle runs faster than the last.
Runs scheduled tasks and chains tools end to end. Self-corrects when a connector goes offline, then reconciles later.
The real shift with AI is upstream of what ships. It changes how the work gets made. So I built the proof on myself instead of putting it on a slide. This is what an AI-native PM looks like in practice: clear stages, real tool orchestration, memory that compounds, and a human in the loop where accuracy matters.
How I Work
This is the operating model I run whether I am leading a product org or inside a short engagement. Six phases, each with a question it has to answer and an output the team can hold me to. It is built to kill weak bets early and get the strong ones in front of real users fast.
What problem is real, and who has it?
Customer interviews, support and sales signal, usage data, and competitive teardowns. AI helps me synthesize hundreds of inputs into themes in hours instead of weeks, then I pressure test the themes with real users.
Output A ranked problem set with evidence behind each one
What are we betting on, and how will we know it worked?
I turn the chosen problem into a crisp bet: target user, outcome, success metric, and the constraints we refuse to break. This is also where I decide what genuinely needs AI and what is better solved deterministically.
Output A one page bet the whole team can repeat back
Does the solution hold up when someone touches it?
Flows and interface design with the team, then a working prototype rather than a deck. I direct AI coding agents to stand up something real in days so we learn from behaviour, not opinions.
Output A tested prototype and a validated or killed bet
What is the smallest version worth shipping?
I slice the bet into releasable increments, sequence them with engineering, and stay in the build daily. Scope moves, dates get defended with tradeoffs, and quality gates are agreed before the first ticket, not after the first bug.
Output A sequenced plan and a build the team can defend
Is the market ready, not just the code?
Launch is a cross functional event. Pricing, packaging, enablement, docs, support readiness, and a staged rollout with a rollback path. Shipped means in customers' hands with telemetry, not merged to main.
Output A release customers can buy, use, and get supported on
Did it move the metric, and what do we do next?
I read the data against the metric we named in Define, talk to the users who adopted and the ones who did not, then decide to double down, fix, or sunset. Nothing stays on the roadmap out of politeness.
Output An evidence based next bet, and one fewer thing to maintain
Compressed into a working sequence when I come in from the outside.
We name the outcome, the user, and the metric that would prove it worked. If a feature does not need an LLM, I say so.
I direct AI coding agents to build something real in days, not a deck. You interact with it before anyone commits a sprint.
Evals against real cases, review gates where a bad output has cost, and a deterministic path for anything regulated.
Auth, data model, observability, and a release pipeline. Shipped means in users' hands with telemetry, not merged.
I leave behind the workflow, not a dependency: how your team scopes, evaluates, and releases AI features on their own.
Decide what actually needs an LLM, and what doesn't.
Reviewable, versioned outputs and approval gates, especially in regulated domains.
Test model and prompt output against real cases before it ships.
Bring AI into how the team researches, specs, prototypes, and ships.
Most AI projects die between demo and production.
The prototype is the easy part. What kills it is undefined evaluation, no owner for bad outputs, and integration work nobody scoped.
AI changes how the work gets made, not just the feature list.
The biggest gains I have seen come from rebuilding the research, spec, and prototype loop, upstream of anything customers see.
In regulated domains, determinism wins.
Let the model produce reviewable artifacts. Let deterministic code run against live data. That is how you get AI leverage without AI risk.
I help teams solve real problems with AI, and ship products that get used.
Two ways to work together, and I'm open about both.
Primary
I'm looking for a senior product role, Director or VP, at a company that wants a practitioner who ships, not a manager who only reviews.
We scope the problem together. If the AI work is not worth doing, I will tell you that.
I take on a small number of engagements: AI product audits, agentic build work, data integration architecture, and fractional product leadership for teams past the AI hype stage.
A clear read on whether your AI bet is real, and what it takes to ship it.
You getA written recommendation with a scoped path to production.
Timeline · 1 week
Let's talkFrom idea to a working AI prototype, fast.
You getA working prototype your team can click, plus the spec behind it.
Timeline · 2 to 3 weeks
Let's talkSenior product horsepower, part-time.
You getA roadmap that ships and a team that can run it without me.
Timeline · Ongoing
Let's talkGo from zero to one by directing AI, the way I built Fairway and Sidekick.
You getYour team building and releasing with agents, end to end.
Timeline · Project-based
Let's talkReferences
Recommendations from clients and leaders I reported to, quoted as written.
It was a privilege to have Josh on my team as VP of Product at Limelight. He successfully navigated highly complex stakeholder environments with creative problem-solving, always delivering beyond expectations even under extreme pressure and budget constraints. I highly recommend him to any company looking for a collaborative, high-impact product leader.
I worked with Josh while he was VP Product at Limelight Platforms, on the strategy behind rebuilding their core SaaS platform from the ground up. What I saw in that work was someone who could set direction and hold it. He made a clear case for retiring the old system rather than patching around it again, then translated that into a roadmap the whole team could execute against.
Every day Josh brought great insight, dedication and a commendable level of professionalism to the very challenging role of Senior Product Manager at Points.com. His mandate to oversee and grow a very large and diverse portfolio of global SaaS offerings was a hard one, which he fulfilled with excellent and unwavering attention to his clients.
Josh is a highly talented, creative project owner. He dove into one of the most complex developments at ADP with enthusiasm. His many years in software development product ownership were a huge asset. Very organized and methodical in his approach to massively complex projects, taking initiative from day one.
I'm open to senior product roles and a small number of AI advisory engagements. If you're building with AI, or working out whether you should, book a free 30 minute call and we'll scope it.
Want a copy of my CV? Send me a message below and I'll reply with one.
Burlington, ON