Early ideas intoworking software.

Working across product and engineering to define, prototype, and ship AI-native products, internal tools, and new ventures.

Built for the messy beginning.

export async function buildPipeline(cfg) {

const agent = new Agent(cfg.model);

const tools = loadTools(cfg.tools);

const mem = new Memory(cfg.memory);

return new Pipeline(agent, tools, mem);

}

pipeline.execute()

.then(result)

.catch(logError);

Name	Status	Last run
Ingest	Complete	2m ago
Ranker	Running	7m ago
Writer	Failed	18m ago

Rough idea
outline
prototype
ship

$ pnpm dev

$ pnpm test

$ pnpm build

✓ all tests passed

✓ shipped

ADOPTION

Black-and-white portrait of Tim O'Connell

ABOUT ME

Hi, I'm Tim O'Connell.

Technical product builder at the intersection of software engineering, entrepreneurship, and product leadership. Comfortable from prototype to production, and from small bets to company scale.

CORE COMPETENCIES

Zero-to-One Product

Find real problems, validate fast, and ship the first version users actually use.

Product Engineering

Build reliable, maintainable software with a bias toward clarity and leverage.

AI-Native Products

Design and ship products that put AI to work in ways that create real value.

Technical Product Leadership

Align people and systems around outcomes and ship with speed and quality.

SELECTED WORK

Things I've Built.

View all projects

SkyTax preview
SkyTax
LIVE
AI-enabled tax prep for international students. From document upload to tax calculation and e-file.
- AI-TOOLS
- SAAS
- FINTECH
Case Study
Keyart preview
KeyartLIVE
Agentic brand management tooling that helps teams maintain consistency across all brand assets.
- AI-AGENTS
- SAAS
- TOOLING
Case Study GitHub
Nightshift preview
Nightshift
LIVE
Toolkit for building and operating agent-based programs. Plan, author, review, build, ship.
- AI-AGENTS
- TYPESCRIPT
- TOOLS
Case Study
Swift Agent preview
Swift AgentBETA
Open-source runtime for streaming, tool-calling AI agents. WebSocket transport, tool routing, and run-level tracing.
- AI-AGENTS
- TYPESCRIPT
- OPEN-SOURCE
Case Study GitHub

HOW I WORK

Principles over process.

A few beliefs that shape how work gets done.

Black-and-white photograph of figures walking along a mountain ridge

Start with the real problem.

Deep understanding before any build. Clarity here saves weeks later.

Scope the smallest useful thing.

Focus is a feature. Small scope, shipped fast, creates momentum.

Build to learn.

Ship early. Get feedback. Iterate with real data, not assumptions.

Use modern tools to move fast.

AI, automation, and pragmatic stacks to ship more with less friction.

NOTES & THINKING

Notes from the wild.

All notes

Why we build in public (sometimes)
APR 28, 2024
From prototype to product
MAR 11, 2024
Hiring like a venture lab
FEB 02, 2024

Let's build something real.

Early-stage projects, internal tools, AI-native products, and technical co-founder or tooling PM/adjacent roles.

Work Together Email

Black-and-white mountain landscape photograph

Rough to real. Fast.