Loading your workspace…
Rounds/Moonshot/Founding Engineer·#90939
Live · accepting submissions
Log in to see accurate information
You're viewing this role as a guest. Sign in to see your application status, referral credits, and personalized match details.
Moonshot logo
AI·Pre-Seed·San Francisco, California

Founding Engineer

at MoonshotAI
Location
San Francisco, California
Salary
$70,000 - $120,000
Type
Full-Time
TL;DR
  • Own iOS surface end-to-end (Swift, on-device)
  • Bar: shipped self-driven GitHub + initiative
  • Primarily front-end/product; real design taste a plus
  • Third brain — problem-solves + pushes back unprompted
About Moonshot
Moonshot builds personal intelligence — the complement to the big labs' general intelligence. Where general models know everything in general and nothing in particular, Moonshot builds a model of one life in particular: your people, plans, promises, and preferences, and what you meant by what you said yesterday. Its vision: "a world where nothing you live is lost and nothing you need arrives too late." The team is a scrappy group of ~19-year-olds (YC-backed) who pivoted here after building and getting acquired for their prior product — deliberately choosing the harder, truer product over vanity metrics. Moonshot ships today as a proactive personal agent for iPhone, in early access with a waitlist. It's built around two ancient jobs: Notes keeps the little moments that turn out to be the big ones (the dose a doctor mentions, the number a contractor quotes), and Reminders brings the right answer back at the exact moment it counts — unprompted. It follows the shape of your day and surfaces the next thing before you search for it. Trust is architecture, not a settings page: microphone audio stays on the device, speech is transcribed on-device, and only text/selected context goes to a model when a moment needs understanding, with real deletion controls. Explicitly not a second brain, not a chatbot, not surveillance. The roadmap: prove the loop on iOS today → a wearable that makes capture effortless → health/safety moments where a surfaced answer matters most → on the far horizon, "kept lives." The house line: take your life personally.
Pre-Seed • 1 – 10
Stage & size
AI
Industry
2026
Founded
About This Role

Moonshot is hiring its third engineer — the founding engineer. The metaphor the team uses: the founders are in the jungle with a machete cutting the path; they need someone to lay the bricks and go deeper — fine-tuning the changes they push, owning surfaces end to end, and acting as a genuine third brain who thoughtfully pushes back on product direction. It's primarily front-end/product on the iOS app (real design taste is a big plus), but a true builder who problem-solves independently: when transcript quality drops because the phone is in a pocket, they're the person who thinks through mouth-to-phone distance, investigates LiDAR, and comes back with a solution — unprompted.

The bar is initiative under ambiguity above all else. On a team this small the spec often doesn't exist yet, and every recent miss was an initiative problem, not a skill problem — so the process ends in a paid multi-day trial on real work with deliberate gaps in the brief. The two features that anchor everything: Reminders and sectionalized Notes. Ramp: month 1 — code in users' hands week one, one surface owned end to end, one thing fixed nobody assigned; month 3 — owns one number (the understanding engine's eval score, checkable via the real harness); month 6 — has visibly moved signups or retention, makes product calls in their area without a founder in the room, and helps interview the next hire.

What You'll Own

  • The iOS product surface (Swift, iOS 26, on-device foundation models) — primarily front-end/product, end to end
  • The two core features: Reminders and sectionalized Notes
  • Independent problem-solving on real product issues (e.g., transcript-quality/hardware problems) — figure it out, don't wait to be told
  • One measurable number (the understanding engine's evaluation score) by month 3
  • Real product decisions in their area; being a third brain that pushes back on direction
  • Fixing things adjacent to the task; reporting back "what you did, what the number did, what you'd change"

Requirements

  • Proof of finished, self-driven work — a GitHub with real projects that shipped and got used (not tutorials, not ten abandoned starts). Code quality is read first — an agent reads the repos; judgment lands on the codebase, not the profile.
  • Initiative under ambiguity — sees what needs doing, owns it, does it, reports back without waiting for a ticket
  • Velocity that survives a real codebase — working Swift landed inside the first days, with or without prior iOS
  • At least one shipped native mobile thing as evidence; iOS/on-device ML are tie-breaking nice-to-haves, not gates
  • Uses AI agents well — shapes/edits what the tools produce rather than pasting; comfortable with agent-built projects

Execution and Ownership

  • A relentless executor with product taste and hacker energy; perfectionist only where trust demands it (privacy architecture, validators), fast everywhere else
  • "Wrong by Wednesday over vague all month"
  • Goes past the ask unprompted; flags a wrong instruction before running it; comes back with what happened + what they'd change
  • Willing to grow into the wearable product area (not a one-lane specialist)

Background

  • 1–2 years, explicitly not 5 (five is long enough to get siloed); pre-PMF, so "hire for IQ" — raw building ability with proof outranks everything
  • School/company pedigree is a weak tie-breaker; education mostly noise
  • Ranking: GitHub/projects → product taste → technical experience → consumer/mobile → startup → prior companies → progression → education → years

Backgrounds That Don't Translate (screen out)

  • Big-company engineers who've only worked inside tickets/sprints/review queues (the habits, not the skill)
  • Pure ML researchers with no shipped product (model work is measured inside a consumer app, not papers)
  • Hackathon-only portfolios (prove starting, not finishing)
  • Agency/outsourcing work (trains building to spec; this role has no spec)
  • Behavioral: best work was assigned to them; serial starters with no finishes

Who Will Thrive Here

  • A builder who's already shipped things nobody asked them to ship, and comes back with what happened, not a status update
  • A relentless executor with product taste who thrives when the brief has gaps
  • Dangerous in Swift fast; happy owning a surface and a number end to end
  • Excited by proactive, on-device personal intelligence — and the wearable future
  • Takes the work personally
Job Details
Experience
0-2 Years
Salary
$70,000 - $120,000
Equity
1.5% - 5%
Visa Sponsorship
Limited
Employment Type
Full-Time
Work Arrangement
In office
Work Intensity
9-9-6
Benefits & Perks
Housing Covered
Food / Daily Necessities Covered
Tools / Company Card For What They Need
Green Flags
Shipped, used personal projects — especially building in public with links
Open-source work on on-device inference / speech / cost-efficiency (batching, etc.)
Indie iOS / SwiftUI shipping; a native app with real users
Red Flags
Big-company engineer who's only worked inside tickets/sprints/queues
Pure ML researcher with no shipped product
Hackathon-only (starts, no finishes); serial starter with no finishes

Required Candidate Q&A

Question 1
Are you able to work in person in SF (strongly preferred), and what's your work authorization status?

Candidate scorecard

· 6 criteria
Proof of finished, self-driven, shipped-and-used work (GitHub; code quality first)
Initiative under ambiguity — owns and does without a ticket
Velocity in a real codebase; dangerous in Swift within ~2 weeks
Rounds · Confidential to recruiting partners · Last refreshed Sep 7, 2026