AI Fullstack with Harness [4+yrs]

About the Role

We are hiring a Full Stack AI Engineer to build production-grade AI agents for the tax and legal domain, where accuracy and reliability are critical.

You’ll own the end-to-end stack—from Next.js/TypeScript frontend and Python backend to RAG pipelines, AI agents, evaluation, and infrastructure. A key part of the role is building production-grade Agent Harnesses that make agents reliable, observable, and safe.

What You'll Do

• Build and ship full-stack AI features using Next.js, TypeScript, and Python.
• Build and operate AI Agent Harnesses covering tool use, planning/orchestration, memory/state, evaluations, guardrails, retries, and observability.
• Build production RAG systems including chunking, embeddings, vector search, hybrid search, re-ranking, and citation-grounded generation.
• Take AI prototypes to production with evaluation, regression testing, monitoring, tracing, and model/prompt versioning.
• Design scalable systems across databases, cloud infrastructure, model providers, and AI runtimes.
• Work directly with domain experts to translate real-world workflows into production-ready AI agents.
• Use AI coding agents such as Claude Code, Cursor, or Copilot as part of your daily workflow.

What We're Looking For

• 2+ years of professional software engineering experience.
• 1–2+ years building and shipping GenAI applications to production.
• Strong Next.js/TypeScript and Python skills.
• Production experience with RAG or AI agents.
• MANDATORY: Hands-on production experience building or owning an AI Agent Harness / Agent Runtime.
• Experience with LangGraph, LangChain, LlamaIndex, or a comparable agent framework, or a custom agent runtime.
• Understanding of LLM limitations, evaluation, observability, reliability, latency, and cost.
• Production experience with AWS/GCP/Azure, Postgres, and object storage.
• Fluent daily use of AI coding agents.Agent Harness experience is NON-NEGOTIABLE. Candidates with only prompt engineering, LLM API integration, or basic RAG experience without meaningful Agent Harness ownership will not be shortlisted.

Nice to Have

• Experience with LangSmith, Braintrust, Arize, Phoenix, LangFuse, or similar evaluation/observability tools.
• Experience in legal, tax, finance, banking, healthcare, or other regulated domains.
• Experience handling sensitive data and compliance requirements.
• Open-source contributions.

What You Get

• Fully remote — work from anywhere with reasonable team overlap.
• $20/month AI subscription credit for Claude, ChatGPT, Cursor, Copilot, Perplexity, or similar.
• Competitive salary.
• Opportunity to build production AI agents and agent infrastructure for high-stakes, document-intensive workflows.