Stealth-AIQ

About the role

We are hiring a Senior Full Stack Engineer to own major surfaces across backend, data, frontend, and system design in an early stage startup environment. This is a builder role: you will ship production features fast, make pragmatic tradeoffs, and iterate directly with customers. You will work across ingestion pipelines, graph and DAG workflows, LLM driven agent systems, copilot UX, and media generation.

You should be comfortable operating with ambiguity, changing priorities, and incomplete requirements. You will be expected to propose solutions, implement them end to end, and harden them for production.

What you will do

1. Build and ship production SaaS capabilities end to end across backend, data, and frontend.
2. Design and implement LLM powered agent workflows using LangChain and LangGraph, including tool calling, routing, memory patterns, evaluation, and guardrails.
3. Deliver GraphRAG experiences, including retrieval, citation, and copilot interactions inside the product.
4. Build ingestion pipelines for GEDCOM, images, and PDFs, handling real world variance, dedupe, entity resolution, and source classification (primary vs secondary).
5. Own AI and media generation integrations (Claude Sonnet, FAL.ai, Remotion and FFmpeg), including reliability, cost controls, and latency optimization.
6. Build graph visualization and interaction patterns for complex trees and DAGs (GoJS, and evaluate alternatives such as Cytoscape).
7. Lead frontend development using React, Vite, TypeScript, TanStack Router and Query, Tailwind, and shadcn/ui.
8. Lead backend architecture with Python, FastAPI, CopilotKit, and PydanticAI, integrating LangChain and LangGraph where appropriate.
9. Design data models and graph workflows using Neo4j (Neomodel) and PostgreSQL (SQLAlchemy).
10. Improve system reliability and operability: observability, tracing, error recovery, performance tuning, and safe deploys.
11. Drive technical decisions, simplify complexity, set engineering standards, and raise the bar on code quality.

Examples of projects you might own

1. Build discovery and ingestion pipelines that unify searches across APIs, parse structured metadata for import, dedupe entities, and auto tag sources as primary or secondary.
2. Implement verification and correlation engines using LangGraph to compute confidence scores, detect duplicates, flag conflicts, and generate explainable reports inside the copilot experience.
3. Build an agent tool layer with deterministic APIs for graph operations, retrieval, and data normalization, with evaluation harnesses to prevent regressions.

What success looks like (first 3 to 6 months)

1. You independently deliver large, chunked product features end to end and get them into customer hands.
2. You reduce complexity through better boundaries, data contracts, and reusable primitives across agents, ingestion, and graph ops.
3. You improve reliability and developer velocity via observability, stronger testing, evaluation for agent flows, and safer deployments.
4. You influence roadmap and architecture through strong judgment, crisp communication, and customer driven iteration.

Preferred experience

1. 7+ years building and shipping production systems with meaningful ownership.
2. Proven startup execution: shipping in small teams, moving fast, iterating with customers, and handling ambiguity and changing requirements.
3. Strong experience building LLM powered systems using LangChain and or LangGraph, including tool calling, retrieval workflows, memory strategies, and evaluation.
4. Experience with graph data models and performance tuning (Neo4j strongly preferred).
5. Experience with messy real world data ingestion and entity resolution.
6. Strong product and UX sensibility: simplify complex data workflows into usable end user experiences.
7. Strong judgment around interfaces, maintainability, and scope control.

Tech stack

• Backend
  Python, FastAPI, CopilotKit, PydanticAI, AG UI protocol
• Data layer
  Neo4j plus Neomodel, PostgreSQL plus SQLAlchemy, AWS S3
• AI and agent frameworks
  Claude Sonnet 4.5, LangChain, LangGraph, PydanticAI (current), CrewAI (planned)
• AI and media generation
  FAL.ai, Remotion plus FFmpeg, Google Vision or Tesseract (planned)
• Frontend
  React, Vite, TypeScript, TanStack Router, TanStack Query, Tailwind, shadcn/ui, D3 or GoJS or Cytoscape, Auth0
• DevOps and infrastructure
  Docker Compose, AWS ECS and Lambda, GitHub Actions

About the role

We are hiring a Senior Full Stack Engineer to own major surfaces across backend, data, frontend, and system design in an early stage startup environment. This is a builder role: you will ship production features fast, make pragmatic tradeoffs, and iterate directly with customers. You will work across ingestion pipelines, graph and DAG workflows, LLM driven agent systems, copilot UX, and media generation.

You should be comfortable operating with ambiguity, changing priorities, and incomplete requirements. You will be expected to propose solutions, implement them end to end, and harden them for production.

What you will do

1. Build and ship production SaaS capabilities end to end across backend, data, and frontend.
2. Design and implement LLM powered agent workflows using LangChain and LangGraph, including tool calling, routing, memory patterns, evaluation, and guardrails.
3. Deliver GraphRAG experiences, including retrieval, citation, and copilot interactions inside the product.
4. Build ingestion pipelines for GEDCOM, images, and PDFs, handling real world variance, dedupe, entity resolution, and source classification (primary vs secondary).
5. Own AI and media generation integrations (Claude Sonnet, FAL.ai, Remotion and FFmpeg), including reliability, cost controls, and latency optimization.
6. Build graph visualization and interaction patterns for complex trees and DAGs (GoJS, and evaluate alternatives such as Cytoscape).
7. Lead frontend development using React, Vite, TypeScript, TanStack Router and Query, Tailwind, and shadcn/ui.
8. Lead backend architecture with Python, FastAPI, CopilotKit, and PydanticAI, integrating LangChain and LangGraph where appropriate.
9. Design data models and graph workflows using Neo4j (Neomodel) and PostgreSQL (SQLAlchemy).
10. Improve system reliability and operability: observability, tracing, error recovery, performance tuning, and safe deploys.
11. Drive technical decisions, simplify complexity, set engineering standards, and raise the bar on code quality.

Examples of projects you might own

1. Build discovery and ingestion pipelines that unify searches across APIs, parse structured metadata for import, dedupe entities, and auto tag sources as primary or secondary.
2. Implement verification and correlation engines using LangGraph to compute confidence scores, detect duplicates, flag conflicts, and generate explainable reports inside the copilot experience.
3. Build an agent tool layer with deterministic APIs for graph operations, retrieval, and data normalization, with evaluation harnesses to prevent regressions.

What success looks like (first 3 to 6 months)

1. You independently deliver large, chunked product features end to end and get them into customer hands.
2. You reduce complexity through better boundaries, data contracts, and reusable primitives across agents, ingestion, and graph ops.
3. You improve reliability and developer velocity via observability, stronger testing, evaluation for agent flows, and safer deployments.
4. You influence roadmap and architecture through strong judgment, crisp communication, and customer driven iteration.

Preferred experience

1. 7+ years building and shipping production systems with meaningful ownership.
2. Proven startup execution: shipping in small teams, moving fast, iterating with customers, and handling ambiguity and changing requirements.
3. Strong experience building LLM powered systems using LangChain and or LangGraph, including tool calling, retrieval workflows, memory strategies, and evaluation.
4. Experience with graph data models and performance tuning (Neo4j strongly preferred).
5. Experience with messy real world data ingestion and entity resolution.
6. Strong product and UX sensibility: simplify complex data workflows into usable end user experiences.
7. Strong judgment around interfaces, maintainability, and scope control.

Tech stack

• Backend
  Python, FastAPI, CopilotKit, PydanticAI, AG UI protocol
• Data layer
  Neo4j plus Neomodel, PostgreSQL plus SQLAlchemy, AWS S3
• AI and agent frameworks
  Claude Sonnet 4.5, LangChain, LangGraph, PydanticAI (current), CrewAI (planned)
• AI and media generation
  FAL.ai, Remotion plus FFmpeg, Google Vision or Tesseract (planned)
• Frontend
  React, Vite, TypeScript, TanStack Router, TanStack Query, Tailwind, shadcn/ui, D3 or GoJS or Cytoscape, Auth0
• DevOps and infrastructure
  Docker Compose, AWS ECS and Lambda, GitHub Actions