CareerPlanGet AI match score →
💼 Full-time🗓 2026-06-25

Core

Building and shipping production AI agents with context engineering, memory, and tools, integrating payment gateways, and working directly on product decisions from day one.

Role type

Senior IC backend engineer specializing in AI agents

Builds

Production AI agents and payment gateway integrations

Domain

Artificial Intelligence / Backend Engineering

Deliverable

production ML models | product features

Required skills

LLM finetuning (LoRA, QLoRA), AI agent development, context engineering, memory management, tool integration, Node.js, TypeScript, Python, FastAPI, Postgres, Docker, payment gateway integration (Stripe, Razorpay, Cashfree, LemonSqueezy), webhook handling, idempotency, reconciliation

Preferred skills

Evals, agent debugging, prompt regression tests, RAG at scale, side project ownership

Technologies

Node.js, TypeScript, Python, FastAPI, Postgres, Docker, Stripe, Razorpay, Cashfree, LemonSqueezy, Claude Code, Codex, Cursor, Windsurf

Responsibilities

Design and implement AI agents with context engineering and memory; integrate payment gateways end-to-end; debug production agent failures; optimize prompt strategies and finetuning decisions; collaborate directly on product decisions with the founder.

Seniority

Senior, hands-on IC

Rewrite
## About the role You work directly with the founder from day one. You sit inside product decisions. You own what you ship. This is not a role where you wait for tickets. We'll hand you a problem and expect you to run with it. We also expect you to say "I'm stuck" the moment you are. Contract to start. Prove yourself, and it becomes full-time with equity. ## You should have - Touched LLM finetuning. LoRA, QLoRA, or a full run — even a small one. You know when finetuning is worth it and when a good prompt is enough. We don't need you finetuning every day. We need to know you're past the tutorial stage. - Shipped at least one real AI agent. Not a demo. Not a hackathon. Something that ran, had users, and broke in production. We want to hear about the breakage. - Built with context engineering, memory, and tools. You know when to use multiple agents and when one good agent with good tools does the job. - Strong in Node.js + TypeScript OR Python + FastAPI. Comfortable in the other. Postgres and Docker are basics. - Shipped one payment gateway end-to-end. Stripe, Razorpay, Cashfree, or LemonSqueezy. Webhooks, idempotency, reconciliation — the boring parts. - Fluent with Claude Code, Codex, Cursor, or Windsurf. You ship 3x faster because of them. ## Bonus - You've built evals, debugged agents, or written prompt regression tests. - You've worked with RAG or retrieval at real scale. - You run a side project that people actually use. ## Not for you if - You need a ticket to start. - You want stability over growth. - You've never shipped an agent. - You go silent for days (weekends off is fine — silence is not). ## For you if - You've shipped something real with AI agents and want to do it ten more times this year. - You get energy from building, not meetings. - You think about what to build even when nobody asked. - You want to join something from day one, not after it's figured out. ## How to apply Email [email protected] with subject: "I build agents" Send three things: - One agent you shipped. What it does. How you built it. What broke, and what you did about it. - Your GitHub or anything live. Links, not PDFs. - One thing you'd build or fix on 91agents.dev right now — be specific, and give us one sentence on why. That's it. No CV. No cover letter. Just the work.
Sourced via wellfound · Listed on CareerPlan, which tracks 70,000+ jobs from 20+ sources.
Apply on Wellfound ↗