Founding AI Engineer (Python, TypeScript, Node.js) | LLM Agents and AI-Native Platform | San Francisco | Up to USD $300K + Founder Level Equity

"Great shot, kid, that was one in a million!"

(Han Solo, Star Wars: A New Hope)

That is roughly the energy this role is looking for. A small team, a very large target, and engineers who would rather take the shot than wait for instructions.

The company

An AI-native startup in San Francisco has just closed an inception round in the tens of millions from one of the most recognisable venture firms in the world. The founders are not first-timers: their previous company grew to several hundred enterprise customers moving billions of dollars of transactions every month, and this new venture is the sharper, more ambitious sequel. Several of the senior team have come across with them, so the engineering and product leadership already know the customers, the data and the failure modes.

The thesis is simple and audacious. There is a category of global commerce that still runs on email, PDFs, spreadsheets and tribal knowledge, and it is measured in the trillions. Rather than bolting a chatbot onto legacy software, this company is building the AI-native system that does the work itself: LLM agents that reason, search, decide and transact. Because the platform sits in the middle of real transactions, every deal it touches makes it smarter, creating proprietary data and network effects that compound with scale. If it works, the outcome is not a nice SaaS exit; it is a new kind of company at the centre of the physical economy.

The team

A dynamic, friendly and passionate group of like-minded engineers and builders who care about the craft and about each other. No prima donnas here. Everyone talks to customers, everyone ships, and everyone is expected to raise the bar for the person sitting next to them. You will be one of the earliest engineers in the room, with a direct hand in shaping the technical direction, the product and the culture.

The role in one line

Own an entire problem domain end to end, build the AI systems that run it, talk to customers directly, and ship relentlessly.

What you will actually build

The problems here are specific and hard, and the engineering lead cares that candidates can speak to them in detail.

1. Entity resolution at scale. Real-world commercial data is chaos: inconsistent identifiers, duplicate records, thousands of catalogues that describe the same thing differently. You will design the matching, deduplication and canonicalisation systems that turn that chaos into a reliable graph.
2. Retrieval-augmented substitution. When the thing a customer wants is unavailable or overpriced, the system needs to find the right equivalent. That means embeddings, vector search, retrieval quality tuning, and LLM reasoning over structured data where a wrong answer costs real money.
3. Agentic search and pricing discovery. LLM agents that hunt for options, compare prices across many parties and act on what they find. Applied agent engineering with tool use, guardrails and evaluation, in production, not in a notebook.
4. A durable-execution agent platform on distributed infrastructure. This is the backbone. Long-running, multi-step agent workflows must survive failures, retries and partial completion. Experience with durable workflow engines (Temporal or similar), event-driven architectures (Kafka or similar) and fault-tolerant distributed services maps directly onto this work.
5. Large, messy, real-world data. Extraction, transformation and manipulation of documents and records at volume, feeding everything above.

Each engineer operates as what the team calls a "GME": you architect, build and iterate on a whole domain with zero hand-holding, then get on a call with a customer to find out whether it works, and go again.

The tech stack, named

Python, TypeScript and Node.js. You need deep production experience in at least one; covering all three is ideal. Around that core: distributed systems, durable execution, complex system design, and hands-on LLM application work (RAG, agents, embeddings, evaluation). No prior exposure to the industry is required. Strong engineers from any domain are welcome.

Top three technical skills, in order of weight

1. Distributed systems and durable execution. Fault-tolerant, scalable services, event-driven design, orchestration that handles failure gracefully. This is the single most important signal.
2. Production fluency in Python and TypeScript/Node.js. Backend systems shipped at a company where the bar was high, not scripting.
3. Applied LLM engineering. RAG pipelines, embedding-based matching, agentic workflows with tool use, and a clear understanding of where models fail and how to contain it.

Seniority: two profiles

The team will hire two archetypes. The first is the early-career engineer (two to five years) with exceptional computer science fundamentals, a strong brand on the CV, and visible rapid progression. The second is the experienced systems architect (up to fifteen years) with deep design expertise who wants founder-level ownership rather than another promotion cycle. A Staff engineer from a large, excellent company who is tired of the pace is squarely in scope.

Where the ideal candidate comes from

The engineering lead has named frontier AI labs, data infrastructure companies and fast-growing AI-native product companies as the calibre of brand they want to see: think OpenAI, Anthropic, Databricks, Snowflake, Confluent, and newer AI companies with serious engineering reputations. Prior titles that resonate strongly: Founding Engineer, Founding AI Engineer, Member of Technical Staff, or Forward Deployed Engineer (Palantir, Scale AI and similar). Early-stage experience on top of a strong brand is the combination the team keeps pointing to.

On education, nothing is mandated, but "exceptional school" appears repeatedly in how the leadership describes their best candidates. Read that as Stanford, Berkeley, MIT, CMU, Waterloo and comparable computer science programmes, ideally with something remarkable attached: a competition result, published work, an open-source project that got noticed, a startup built while still studying.

The non-technical bar

Three traits are weighted as heavily as the code.

High slope. Evidence of fast, continuous growth: rapid promotion at a strong company, then a jump to founder-level ownership somewhere smaller.

Autonomy and hunger. The leadership says it plainly. Show up early, stay late, run at hard problems without being pointed at them, and ship.

Customer-facing comfort. This is a dealbreaker, not a nice-to-have. Every engineer will speak with customers and partners directly, so commercial savvy, product sense and the personability to hold a conversation with a buyer matter.

They also want one remarkable thing. A standout achievement in school or work that signals top-tier.

Logistics

San Francisco, five days a week in the office. Base salary from USD $150K to $300K with founder-level equity on top; the equity is described as a central part of the package, not a token. The interview process is four steps: an initial conversation, a live-coding technical screen, then an onsite covering technical depth, a product case study and a deep dive into a project you owned. The team moves fast and has made offers within hours for the right person.

What it is not

It is not remote, and it is not hybrid. It is not a role for someone who wants a product manager to write the ticket first. It is not a research or pure modelling role; the work is systems engineering with LLMs as a core component. And it is not a fit for profiles built on part-time or remote consulting gigs dressed up as startup experience, or for CVs with three consecutive sub-one-year stints and no clear upward trajectory. The leadership has been explicit on both.

If this is you

You have owned a large distributed system end to end, you write Python and TypeScript at a professional level, you have shipped something real with LLMs, you can sit across from a customer and explain what you built, and you have one story that makes people lean in. Take the shot.

Visas and relocation

Yes to both sponsorship and transfers. The company will sponsor new visas (new H-1B and TN) and will accept transfers of existing status (OPT and H-1B transfers). Relocation to San Francisco is supported for candidates coming from elsewhere in the US or overseas. If you are on OPT or STEM OPT, or currently holding an H-1B with another employer, this is a live option.

Founding AI Engineer (Python, TypeScript, Node.js) | LLM Agents and AI-Native Platform | San Francisco | Up to USD $300K + Founder Level Equity