Keplaris is a New York product engineering studio, founded in 2025 and remote-first across the US. We build our own products — GeoIPHub, ClickFortify, and Tether — and deliver client work in e-commerce, web, APIs, and automation. Running our own software keeps us honest: we build automation for clients the way we'd build it for ourselves, because we know what breaks at 3 a.m.
Our automation practice has one rule: the system has to actually work after we leave. Not a demo, not a brittle chain of zaps — a production system with error handling, logging, and a human in the loop wherever judgment matters.
The team is small and technical. You'll work alongside full-stack engineers who run real APIs in production, which means your automations get built on solid infrastructure instead of duct tape.
This role matters because most AI automation sold today is theater. We want the engineer who can tell the difference — who knows when an LLM call genuinely solves the problem, when a cron job and a webhook do it better, and how to wire either one so a business can depend on it.
What we look for
- LLM pragmatism. You've built with the OpenAI and Claude APIs in production — prompting, structured outputs, tool use, evals — and you reach for a model only when it beats deterministic code.
- Integration chops. You connect CRMs, e-commerce platforms, spreadsheets, and internal tools through their APIs, and you handle auth, rate limits, and retries like the adult in the room.
- Human-in-the-loop design. You build systems that know their own limits — routing low-confidence decisions to a person, making approvals fast, and keeping an audit trail.
- Reliability mindset. Your workflows log what they did, alert when they fail, and recover without someone re-running things by hand. Silent failure is the enemy.
- Business translation. You can sit with a client, find the process that's eating their week, and turn it into a system — then explain what it does in language they trust.
- Code first. You're a real engineer — comfortable in TypeScript or Python — who uses workflow tools when they fit and writes custom services when they don't.
What you'll do
- Design and build automation and AI systems for clients — document processing, support triage, data pipelines, and operations workflows with humans in the loop where it counts.
- Build LLM-powered features into client products and our own, with proper evaluation instead of vibes.
- Connect business tools — Shopify, CRMs, email, billing, internal databases — into systems that run unattended and fail loudly.
- Own delivery end to end: scoping with the client, building, deploying, monitoring, and handing over something they can actually operate.
- Help define the Keplaris automation playbook — the patterns, guardrails, and standards behind every system we ship.
What we offer
A remote-first US team, competitive pay, and real product ownership — the systems you build run in production for businesses that depend on them, and you own them from first call to handoff. You'll work with engineers who run their own APIs and hold automation to the same standard, with direct access to founders and no process for the sake of process.
How to apply
Email main@keplaris.com with your resume and a GitHub or portfolio link. Tell us about one automation or LLM system you built that ran in production — what it replaced, where the human stayed in the loop, and how you knew it was working.
Application form
Apply for the Automation & AI Engineerrole — tell us about yourself and we'll be in touch.