Role overview
*Technical Lead / Software Architect -- Generative AI Hyperpersonalization Platform
$96,000-$120,000/year | Remote (US-based preferred) | Full-Time Contractor | 0.5-1.0% Equity**
HeartStamp is an AI-native hyperpersonalization platform for greeting cards, invitations, and gifting. Users create deeply personal, professionally printed physical products through a guided AI experience. We generate production-ready, print-perfect output (PDF/X, 300 DPI CMYK) fulfilled through our print-on-demand partner network. MVP launches April 2026. Scaling across the US, UK, and Canada.
We're a bootstrapped, founder-led team of 25+ across engineering, product, design, and growth. We move fast, we ship, and we build to win.
THE ROLE
We need a Tech Lead who owns the technical vision and stays deeply hands-on in the codebase. This is a player-coach role -- expect 60-70% of your time writing, reviewing, and debugging production code. You set architecture direction and lead the engineering team, but you build alongside them every day.
This role is heavily focused on AI/RAG systems, semantic search infrastructure, image generation pipelines, and DevOps/infrastructure strategy. You'll lead a team of full-stack engineers, a RAG engineer, an AI orchestration engineer, a DevOps engineer, and a QA lead.
What you'll work on
Own the end-to-end technical architecture: Python backend, Next.js frontend, AI/RAG pipelines, vector search infrastructure, and AWS cloud platform.
Lead development of our RAG pipeline using PostgreSQL/pgvector, LangChain, LangGraph, and LangSmith. Improve hallucination handling, prompt engineering, content validation, and multi-model inference routing across a model-agnostic architecture.
Take direct ownership of AWS infrastructure strategy, cost optimization, autoscaling, and production reliability. Architect for zero-floor scaling with no idle burn. Manage Kubernetes clusters efficiently and maintain CI/CD pipelines.
Write production code daily. Review code. Debug production issues. Harness the full potential of Claude Opus 4.6 and Codex 5.3 to move at a velocity that would require three engineers without AI acceleration.
Run sprint planning, code reviews, architecture discussions, and daily standups for a distributed engineering team across multiple time zones.
Ensure our generation pipeline produces print-perfect results: correct color profiles, bleed zones, resolution, and format compliance through our DocRaptor/PrinceXML print pipeline.
Build for international expansion (multi-currency, multi-language, multi-partner routing), product extensions (digital 3D cards, invitations, gifting), and a future creator marketplace with royalty structures via Stripe Connect.
What we're looking for
- 6+ years of engineering experience designing and building complex, scalable production systems
- Deep backend expertise in Python/FastAPI (our backend is Python -- you need to be authoritative here)
- Real RAG and vector search experience: pgvector or comparable vector databases, embedding strategies, retrieval optimization, LangChain, LangGraph, LangSmith
- Full-stack capability in React/Next.js and TypeScript
- Strong DevOps and infrastructure experience on AWS with Kubernetes, GitHub Actions CI/CD, containerized microservices, Terraform/IaC, and a track record of infrastructure cost optimization
- AI-native workflow using Claude, Codex, Cursor, or similar tools as daily force multipliers
- Current, hands-on coding proficiency (we will evaluate this in our interview process)
- QA-driven engineering mindset
- Strong written communication and experience leading distributed async teams