Colin GillinghamBook a call

Colin Gillingham · New York · 2008 to now

Reverse
engineer.

18 years6 disciplines20+ products since 2023

Self-taught across sales, GTM ops, growth and marketing, product and engineering, and now AI. Eighteen years of taking the machine apart to learn how every part works.

18 years2008 to now

28 moments across six disciplines. Press play to walk through them in order, or click any dot.

28 moments

2008Now

Each ring is a year, 2008 at the center
Shaded: a year I worked in that discipline
A moment. Click one, or press play
A specialist’s rings, for scale: one slice

AI products I built

Selected work

Six AI products I designed and built myself since 2023, and what’s new about each one. Flip through them here, or take the PDF with you.

Download PDFFull screen ↗

Selected AI work · 2023–2026

Colin
Gillingham

Reverse engineer.

SalesGTM OpsGrowth & MarketingProductEngineeringAI
01 / 09Cover

AISalesEngineering

PhoneScreen

2024–now · My company · Live

Voice agents that run the first phone interview for hourly jobs, at any hour. I built the whole pipeline from scratch on OpenAI and Twilio.

2,500+
interviews completed
80+
hiring teams
  1. Candidate applies From Indeed, CareerPlug and other job boards
  2. Agent calls A voice agent on OpenAI realtime and Twilio
  3. Interview runs Day or night. 58% happen outside 9 to 5
  4. Results reach the team In the hiring tools they already use

ProductAIEngineering

Glottie

2026 · App Store

An iPhone language tutor you learn from by talking. Underneath, it’s a language-learning harness I built around OpenAI’s realtime voice model.

1,000+
language combinations
63 days
first commit to the App Store
  1. Pick any pair 41 languages, 1,000+ combinations
  2. Talk Out loud, no typing
  3. The harness steers It keeps the realtime model on lesson
  4. The tutor answers Live, in the language you’re learning

AIEngineering

Agent Wonderland

2026 · Live

A marketplace where AI agents find, pay for and run other agents. I built the MCP discovery layer and the marketplace, with payments on x402 rails.

9,600+
npm downloads
x402
payment rails
  1. An agent needs help Say, a filings analysis in Claude Code
  2. It finds a specialist Through my MCP discovery layer
  3. It pays per request Over x402
  4. It runs the job Results come straight back

AIEngineering

TinyNews

2023, 2026 · Live

A news aggregator that reads thousands of articles a day and gives every reader a feed of their own.

1,000s
articles indexed a day
4 days
first commit to launch
  1. Index Thousands of articles, every day
  2. Classify A lightweight classifier keeps it fast and cheap
  3. Tailor A bespoke feed for each reader
  4. Next Your newsletters, pulled in by email
UXLoops: a clickable prototype with the research agent beside it.

ProductAI

UXLoops

2026 · Live

Customer research without the calendar. Send a customer a prototype, and an agent runs the research while they explore it.

Async
no sessions to schedule
  1. Prototype Describe a product, get a clickable prototype
  2. Customer explores On their own time
  3. Agent interviews The questions a researcher would ask
  4. Their version Feedback becomes the customer’s view of the prototype

SalesAIGrowth & Marketing

Peregrine

2026 · Launching now

Speed to lead for B2B teams. A voice AI that calls every new lead seconds after the form and live-transfers the good ones to a free rep.

4 sec
form to first call
29 hrs
B2B average (RevenueHero, 2024)
  1. Describe a good lead Peregrine drafts the questions; you approve them
  2. Called in seconds It looks up the company before it dials
  3. Qualified on the call It asks only what it can’t look up
  4. Live transfer To the next free rep, while the lead is on the line
01 / 06PhoneScreen

Consulting · AI assessment

Five days with your team.
A clear plan for AI.

I work inside your team for a week, find where AI will pay off, and leave you with a ranked plan and a working demo. If we build it together, the fee comes off the project.

  1. Day 1InterviewsHow the work actually gets done, from the people doing it.
  2. Day 2WorkflowsWhere the hours, handoffs and copy-paste go.
  3. Day 3SystemsYour data, tools and owners, mapped end to end.
  4. Day 4PrototypeA working demo of the task your team hates most.
  5. Day 5ReadoutA ranked plan: what to build first, and what it’s worth.
$1,500 · five days
ThenBuild itFixed price, shipped in four weeks or less, in your repo.
Build, from
$7,500
Assessment credit
−$1,500
You pay, from
$6,000
Book a 30-minute callcolingillingham@gmail.comOn site in New York, or remote

Contact

Two ways to work together

HiringAI product leadership, in New York.Group PM, Director or Head of Product on AI products. On site or hybrid.colingillingham@gmail.comConsultingStart with the five‑day assessment.$1,500, credited in full toward the build. On site in New York, or remote.Book a 30-minute call