AI Product Manager

Logan Kilpatrick Portfolio

Works on Google AI Studio and the Gemini API at Google DeepMind, after leading developer relations at OpenAI from 2022 to early 2024; earlier built lunar-rover mission-planning software at NASA and trained ML models at Apple. Has made over 50 AI startup angel investments, including Cursor and Cognition.

Developer ToolsGemini APIGoogle AI StudioDeveloper Relations

Location

Coordinates unknown. Vibes: international.

Skills

Developer ToolsGemini APIGoogle AI StudioDeveloper Relations

What makes it work

A breakdown of the choices that make this portfolio stand out.

1

A career sequence from safety-critical software to consumer developer tools

Going from NASA lunar-rover mission planning to Apple ML training to OpenAI developer relations to Google AI Studio product leadership shows a specific throughline across very different stakes and audiences.

2

Naming the specific transition from DevRel to product leadership

Explicitly stating he moved from leading developer relations into a product lead role shows a deliberate career move from advocacy work into direct roadmap ownership.

3

Investing as an extension of product judgment, not a separate hobby

Fifty-plus angel investments in AI startups, named specifically, function as a public track record of picking which AI products will matter.

What AI Product Managers can take from this

Specific, actionable tips to apply to your own portfolio — no generic advice.

  • If your career moved between very different domains, name the throughline explicitly rather than listing roles chronologically with no connective explanation.

  • Distinguish clearly between developer-relations/advocacy roles and product-ownership roles — a reader shouldn't have to guess which one you held.

  • If you invest in or advise startups in your space, name a few specific companies — it's a form of applied judgment that's otherwise hard to demonstrate.

Case studies

Logan Kilpatrick's personal site is a bio, a list of beliefs and an investment list. The work lives on other people's domains: Google's developer blog, a Gemini starter repo and a podcast feed. The cards collect it from there.

Case StudybeginnerEpisodes through July 2025

Around the Prompt

PodcastLong-form interviewsApple PodcastsSpotify

An interview podcast he co-hosts with Nolan Fortman: roughly hour-long 'unfiltered conversations' with AI founders, researchers and investors about how the field is changing.

Regular long-form conversations with founders building on these models keep a platform PM close to what developers ship on top of his product.

  • Bolt.new episode covers its growth from $0 to $40M ARR

  • Mix of coding-tool, robotics (Figure, Matic) and investor (a16z's Anjney Midha) guests

  • Latest listed episode: Factory AI's Matan Grinberg, July 2025

Projects

SaaS ProductadvancedLaunch posts from June 2024 to May 2026

Google AI Studio and the Gemini API

Gemini APIGoogle AI StudioContext cachingCode executionManaged Agents

Google's developer surface for Gemini: a browser workspace for prompting and building apps, plus the API behind it. Kilpatrick co-authored the launch posts for major updates, e.g. opening the 2M-token context window to all developers with context caching and code execution (June 2024), and Managed Agents plus native Android app building (I/O, May 2026).

Launch posts show the job: pairing a model capability with the pricing and tooling around it, e.g. context caching shipped alongside long context to cut repeat-token costs.

  • Context caching launched to offset the input cost of 2M-token prompts

  • Code execution runs in a sandbox with no internet; billed on output tokens only

  • I/O 2026: Managed Agents spin up a tool-using agent in an isolated Linux environment with one API call

Open SourcebeginnerPublic repo on his GitHub

Gemini API Quickstart (Python)

PythonFlaskGemini API

A small Flask app with a basic UI that gets a developer from zero to a working multimodal Gemini call. The README walks through a first chat request and shows how the chat object's history changes.

Developer-product work includes a short path to first success. A 'running in under 5 minutes' starter is that path, shipped as code.

  • Pitched as 'up and running in under 5 minutes'

  • Flask backend plus basic UI, not just a script

  • Maintained under his own account, pinned beside the official cookbook

Ready to build your portfolio?

Follow the AI Product Manager roadmap — skills, projects, and timeline to get hired.

AI Product Manager Roadmap