Marina Zorina
AI-native product builder · Founder of CrewTask
Summary
I take a vague idea to a shipped, reliable AI product and own it end to end - CrewTask, live on Google Play, built solo. I own what gets built and how we prove it works: strategy, architecture decisions, decomposition, acceptance criteria - and I run a team of AI agents the way a product owner runs a team. Fifteen years in engineering - on why systems fail and how to prove they don't - is where the reliability comes from.
Looking for: Co-founder & CEO who builds, Head of Product / AI Product Lead or Entrepreneur in Residence at an AI-first startup or venture studio, pre-seed to Series A.
Products I Built
- Took the product from idea to a live Android app in Google Play open testing, solo and part-time: about 3 months of development plus a month of store release.
- Voice note in, structured family tasks out: three-tier Gemini orchestration with a Groq fallback; an eval suite of 49 labeled examples in three languages, with an accuracy threshold checked before parser releases.
- Multi-tenant Supabase with RLS, offline-first sync, 330 unit and 96 E2E tests, Sentry in production. Next: M-Tribe, a community layer of verified families nearby.
- Slash commands and checks that make an AI coding agent show its work instead of just saying "done", plus an 8-module release pipeline distilled from CrewTask's production.
- 14 commands and 21 prompts, each scored: structure evals in CI and an LLM judge over written cases.
- 91 of 102 commits, all 26 pull requests and all 57 issues in two weeks: engine work on Android and iOS, assertions, conditional steps, per-step JUnit report and CI.
Also built solo through AI orchestration: Health Assistant (MVP with a live demo; on-device document recognition and a state model where every number is calculated by fixed formulas in code, and the LLM only explains the results in plain language) and a Research Agent (Python, 13 login-free sources, three-layer LLM matching, running twice a day).
Client Work
- Help shape how the product is released and work on the product itself: parts of the code, the testing framework, and a UI kit with a review of the product's screens.
- Found green reports standing over runs whose checks never executed and closed the gap in CI; the daily report grew from 972 checks to 7,479. Regulated industry: synthetic personas only.
- API-heavy product team: payments, fraud detection, account linking.
Background
- Built the release function at an EdTech from scratch: hired and led a team of 3; release cycles went from 1-2 weeks to 1-2 days.
- What I built kept working after I left: the two people I grew both moved on at about twice their salary, and the automated reports kept running for years.
- Business analysis and budget management for manufacturing operations.
- OLAP management reporting from invoice to C-level KPIs; multi-year financial modeling for the unbundling of RAO UES, the restructuring of Russia's national electricity grid.
Skills
- Product
- 0-to-1 strategy, architecture decisions, decomposition, acceptance criteria, release ownership, evals and guardrails for LLM features
- AI
- Multi-agent orchestration, context engineering, agentic workflows, MCP; Claude, Gemini (flash-lite / flash / pro), Groq, multi-provider pipelines
- Stack
- React 19, TypeScript, Supabase (Postgres + RLS), Deno Edge Functions, Capacitor, Tailwind, Python
Looking For
- Role
- Co-founder & CEO who builds · Head of Product / AI Product Lead · Entrepreneur in Residence
- Stage
- Pre-seed to Series A, AI-first startup or venture studio
- Format
- Full-remote from Mexico. Preferred, not required: GMT-5 ± 4 hours
Education & Personal
- Mathematics & Economics, Honors - Vyatka State University
- Second degree: instructor of applied combat sports - Russian State University of Physical Culture
- Sport since early childhood: martial arts (MMA, grappling), swimming, trail running, diving · two kids
- English, Russian · Spanish in progress
Contact