Member of Technical Staff

Lightsage; San Francisco (CA); On-site; USD 140,000 - 210,000; Full time

Job summary

Lightsage is hiring an early-stage engineer in San Francisco to own systems end-to-end across frontend, backend, infrastructure and data. You’ll work directly with customers to turn problems into shipped features and contribute to rapid product iteration with AI coding tools.

As one of the first engineers, you’ll shape architecture, collaborate with the founding team, and build scalable solutions that power enterprise integrations. This demanding, on-site role offers strong ownership and growth.

Qualifications

• Track record of taking a loosely defined customer or product problem to a shipped solution with little direction.
• Strong backend experience building reliable systems that handle production load.
• Hands-on experience shipping quickly with AI coding tools such as Claude Code or Codex.
• Clear thinking about technical trade-offs, and clear written communication.
• Experience doing research or running rigorous experiments, such as LLM evals, benchmarks or A/B tests, and drawing sound conclusions from noisy results is a plus.
• Experience with Next.js, FastAPI, Supabase or Microsoft Azure is a plus.
• Prior founder or early-stage startup experience is a plus.

Responsibilities

• Own end-to-end systems across frontend, backend, infrastructure and data.
• Work directly with customers to translate their problems into shipped features.
• Develop and run experiments/benchmarks to measure impact on guidance quality.
• Build agents and benchmarks to improve recommendations and reliability.
• Scale evaluation and data pipelines for enterprise customers.

Skills

Backend systems

AI tooling usage

Clear written communication

Experimentation & A/B testing

Customer collaboration

Tools

Next.js

FastAPI

Supabase

Microsoft Azure

Job description

Company Description

AI agents are becoming how software gets chosen. Developers ask Claude Code, Codex, Cursor and ChatGPT which tool to use, and more and more they let the agent pick and integrate it. If an agent doesn't recommend you, or loops trying to figure out how to use your product, it moves on to something else.

Lightsage gets companies ready for the agent economy: visible to agents, so they get recommended, and usable by agents, so integrations work the first time. We track how often coding agents and AI answer engines recommend a company against its competitors, run real coding agents against its APIs, CLIs and MCP servers to find where they fail, and tell the team exactly what to change.

We work with category leaders in AI and developer tools, including Firecrawl, Reducto, Scalekit, TinyFish, Daytona, Rime and Meetstream, and we run DevToolArena.com, a public benchmark that ranks developer tools by how well coding agents use them.

We've raised a $4M seed led by Nexus Venture Partners, and we're backed by growth leaders and founders of Postman, Apollo, Firecrawl, Resend, Daytona, TinyFish and others. Our goal is to be the default for any company that wants to win with agents.

Role Description

This is one of the first engineering roles at Lightsage. You'll join the founding team in San Francisco. You'll own systems end to end across frontend, backend, infrastructure and data, and work directly with customers to turn their problems into shipped features.

Early projects include:

• Improving recommendation quality, and measuring whether a recommended change actually helps a customer
• Designing the experiments and benchmarks that tell us which changes actually make agents recommend and use a product
• Building agents that write, publish and re-test documentation changes for customers
• Adding support for new coding agents and models as they're released
• Scaling eval and data pipelines for enterprise customers

New models ship every few weeks, so the work changes constantly. The team builds with AI coding tools and our own agent orchestrator, and reviews all code before it ships.

Requirements

• Track record of taking a loosely defined customer or product problem to a shipped solution with little direction
• Strong backend experience, including building reliable systems that handle production load
• Hands-on experience shipping quickly with AI coding tools such as Claude Code or Codex
• Clear thinking about technical trade-offs, and clear written communication
• Experience doing research or running rigorous experiments, such as LLM evals, benchmarks or A/B tests, and drawing sound conclusions from noisy results is a plus
• Experience with Next.js, FastAPI, Supabase or Microsoft Azure is a plus
• Prior founder or early-stage startup experience is a plus

This is a full-time, on-site role in San Francisco. It's a demanding schedule that regularly includes weekends, with significant ownership over the company's technical direction.