Staff Engineer - San Francisco

About Eli

America is investing trillions of dollars to make homes and buildings more efficient, electric, and resilient. But the programs paying for these upgrades still run on paperwork, manual review, and disconnected systems. Contractors can wait months to receive money they have already earned. Program administrators struggle to process projects quickly and accurately. Too much funding is lost to administrative friction before it reaches the buildings it was intended to improve.

Eli is building a better system.

Our platform helps utilities, government agencies, and their partners run energy programs, from determining whether a project qualifies to reviewing documentation, reporting results, and issuing payments. We also advance money to contractors before the underlying rebates are paid, giving them the cash flow to keep taking on new work.

Since launching in December 2023, Eli has processed hundreds of thousands of projects and advanced nearly $50 million to contractors. We are now preparing to support a much larger volume of projects and capital.

The Opportunity

We're hiring a Staff Engineer. Under the hood, Eli takes in applications from contractors and homeowners, verifies eligibility and documents, decides who gets paid, pays contractors early through a credit facility, tracks every dollar through a ledger, and reconciles it against the bank. A lot of what we build moves real money and gets audited.

You'll work directly with our two cofounders: Adrian, who heads product, and Eduardo, who heads engineering. We're a team of 7 engineers, including Adrian and Eduardo, both of whom commit to production daily. Each engineer owns a large part of the system, and your work shows up in production and in front of customers within days.

As a Staff Engineer, you'll set technical direction for the systems where correctness matters most: money movement, AI-driven decisions, and the platform underneath them. You'll still write and ship code every week. Here, seniority means more responsibility for outcomes, not more distance from the work.

This role is based in San Francisco, and we work together in person there. For the right engineer, we can accommodate fully remote within the US. Because of compliance requirements, this role is open to US residents only.

How We Work

Almost all of the code we ship is written by AI agents. We run Claude Code across parallel worktrees, and our own bots handle PR review and first-pass production investigations. Our hand-crafted harness and internal tooling make this possible. We review everything carefully, and the engineer whose name is on the PR is responsible for every line in it.

Good judgment is the scarce resource. An agent will write the pagination code. It won't notice that the query loads 1,000 rows to render 50, question whether the feature should exist, or explain to a program manager why a claim was denied. That part is the job.

What You'll Own

Nobody here will groom a ticket for you and wait. We want someone who looks around, sees the work nobody is doing, and takes it. Some of what's on the table today:

• Technical direction. Decide how the ledger, decisioning, and platform evolve, write the designs other engineers build on, and raise the bar through code review and pairing.

• Product, end to end. You'll figure out what processors, contractors, and program administrators need, build it, watch it work or fail in production, and adjust.

• AI decisions that move money. Document verification, eligibility checks, and fraud detection, where a wrong answer means someone gets paid who shouldn't, or doesn't get paid when they should. You'll help decide when to use a model, a rule, or a human, and how to measure whether it's working.

• Money movement. The ledger, the loan tape we report to our lender, and reconciliation against bank records. This has to be correct to the cent and hold up in an audit.

• Our internal tool stack. Consolidating the systems we've outgrown, like Airtable, into our own agentic, in-house tools for operations and program configuration.

• The platform and production. CI speed, test-suite health, backups, database migrations, alerts, and workflows. When something breaks, you find the root cause, fix it, and write down what happened.

What We're Looking For

• You're a strong engineer with the AI turned off. You read the error message before anything else, debug from first principles, know git well and can't imagine working without rebase, and can reason through a request from the browser to the database and back.

• You use AI agents heavily and remain the author. When someone asks you a question in a meeting or on Slack, you answer from your own understanding. You can defend any line of your PR and say how you checked that it works.

• You measure your week by problems solved, not tickets closed.

• You care about the business. You'll learn how rebate programs, underwriting, and contractor payments work, and you'll push back when a spec is wrong.

• You're comfortable and competent in compliance- and security-heavy environments. We're SOC 2 audited, we handle personal and financial data, and the programs we serve have their own rules and auditors.

• You write short, clear updates: what's done, what's blocked, and when the next thing lands.

• You have strong opinions about your tools and can tell us why.

Experience:

• 8+ years building production software (ideally some of it before AI coding tools), including time as the person responsible for a system when it breaks.

• Staff-level scope or equivalent impact: you've led the design of a system other engineers build on and stayed accountable for it in production.

• Deep TypeScript on the backend (Bun, Node, Next.js, tRPC, Prisma or similar) and solid relational database skills in MySQL or Postgres: schema design, query plans, migrations.

• A track record of taking on work outside your assigned lane and finishing it.

Nice to have:

• Temporal or another workflow engine

• Payments, ledgers, lending, or reconciliation

• dbt, Snowflake, or other analytics tooling

• Evaluating LLM output with test sets and metrics

• Infrastructure: Pulumi, Docker, Vercel, Tailscale, Datadog, Sentry

• Energy programs or utility software

Stack: TypeScript Turborepo monorepo (pnpm with tsx for older systems, Bun for newer ones) · Next.js, React, tRPC, Prisma · MySQL, moving to Postgres · Temporal · Python for document AI · Snowflake, dbt, Lightdash · Vercel, DigitalOcean, Pulumi · Datadog, Sentry · Claude Code with worktrees, plus internal agent tooling

You'll help shape how the engineering team grows, including who we hire next.

Compensation and Benefits

The base salary range for this role is $220,000 - $250,000 annually, depending on experience, plus a meaningful equity package.

• Generous medical, dental, and vision coverage

• Flexible, unlimited time off, with a minimum of three weeks encouraged

• At least two company gatherings each year and budget to co-work when solving difficult problems together

• A welcoming, family- and parent-friendly environment that values humility, honesty, and humor

Eli has a no-negotiation compensation policy. We believe it is our responsibility to understand your expectations during the interview process and make the strongest offer we can. If you receive an offer from Eli, it will be our best offer.

We encourage you to apply even if your experience does not map perfectly to every bullet.