For over a decade, my world has been defined by scope management, vendor relations across 460 syndicated channels, and driving the end-to-end delivery of complex digital products for massive brands like PlayStation and DC Universe. Now, I’m taking all that experience and focusing it on a new passion: building apps with Google Gemini Pro and AI Studio.
You might think switching from a Lead Project Manager role on a $10M+ streaming service campaign to coding an AI app is a huge leap, but honestly? It feels more like a seamless transition. My PMP toolkit isn’t just for overseeing a massive team; it’s the secret weapon for my solo dev hobby.
Here’s how the core concepts of project management are translating directly into building my own AI apps:
1. Scope Management: The AI App’s North Star
In PM, we constantly fight scope creep. An unchecked feature request can sink a timeline and a budget. Now, when I start a new AI project in Google AI Studio, the first thing I do is define the AI App’s Minimum Viable Product (MVP).
- PM Mindset: What is the essential problem the project needs to solve? (e.g., Launching the DC Universe originals like Titans or Harley Quinn).
- Dev Mindset: What is the single, core instruction my Gemini model needs to execute perfectly? If I’m building a simple AI consultant for small businesses (like I’ve done in the LA area), I stick to a narrow focus, like analyzing a simple business plan for risk, before adding in a feature to draft a full content calendar.
This discipline keeps me from getting lost in the almost infinite possibilities of AI.
2. Risk Assessment: Proactive Prompt Engineering
As Project Managers, our job is to see around corners and manage potential business risks. In AI development, this is less about platform partners like Roku or Samsung and more about anticipating how the model might fail, wander off-topic, or even generate unhelpful responses.
- PM Translation: I treat potential model drift or an unhelpful response as a project risk.
- The Solution: I proactively mitigate this risk through aggressive prompt engineering. I set explicit guardrails, define tone and style, and limit the scope of the data the model can access. Just like I established relationships with vendors like AMVD and Akamai for video content delivery, I have to manage the “vendor”—which is now the AI model itself—to ensure reliable output.
3. Iterative Development: SCRUM for One
My professional background includes conducting tactical stand-ups (SCRUM) for Channel Marketplace Ops teams. Agile is all about short, rapid cycles of testing and refinement.
When I’m in AI Studio, I’m running a SCRUM for One:
- Sprint Planning: Define the feature I want to build or the prompt I want to test.
- Daily Stand-up (with myself!): Review what worked, what broke, and what I’ll tackle next. Did the Gemini model accurately summarize partner-related issues ingested via Jira?
- Sprint Review (Testing): Test the app multiple times with different inputs. Tweak the instructions or the configuration based on the output.
This iterative approach—test, fail, fix, repeat—is how I refine the model and quickly deliver an app that actually solves the problem I defined in my scope. It’s the fastest path from idea to functional app, and it’s straight out of the PM playbook!
My journey from overseeing complex implementations to developing my own AI tools (check out 3 of them here) is proof that Project Management isn’t just a job; it’s a superpower you can apply to any passion, even app development!
What is one PM concept you think could instantly make you a better coder? Tell me in the comments! 👇
Leave a Reply