Founding Engineer

About BotBot

BotBot helps teams test and improve how voice AI behaves in live customer conversations, and measure the impact on real business outcomes.

The role

We're hiring a founding engineer to help build BotBot and the system we run the company on. This role is close to the product, customers, and architecture, and you will help decide what we become.

BotBot is the interaction intelligence system where teams design AI behavior, test it in live experiences, and measure the result. Underneath it, we are building our operating model into software: shared context and goals become agent work, and every piece of that work is checked against what the company believes. We run on that system every day, and it is becoming part of what we ship.

We removed the junior/senior divide from this role on purpose. We are open to the right person at any point in their career; the mission scales to the person, and the expectations below describe how you work, not how long you have worked.

You will receive missions and company context rather than a maintained backlog. We expect you to determine what work would move the mission, decompose it into parallel human and agent work, and keep the queue moving until the outcome is shipped or a consequential decision needs input.

What you'll work on

This is a generalist engineering role. You will work alongside a founding design engineer who owns the depth of the product's interfaces; you own whatever layers are required to make the product real and trustworthy, and the immediate weight is on the systems underneath.

Concretely, you might:

• build the runtime that turns goals and shared context into agent work, verification, and review
• design the APIs, data models, analytics, SDKs, and integrations behind the product
• build experiment workflows that connect behavior changes to business outcomes
• build trust into the product through clear state, audit history, verification, rollback, and honest handling of missing information
• implement product surfaces end to end when that is what the mission requires
• create fast feedback loops using generated variants, browser automation, and comparative review

You will also build and improve your own execution system: spec pipelines, agent workflows, verification, visual testing, and the tooling that makes each project faster than the last. The work is complete when the parts have been integrated, tested, reviewed at the right level, and shipped as one coherent product.

How you operate

You should be comfortable taking a goal such as “make agent review understandable to a founder” or “connect this behavior change to a business outcome we can defend” and creating the work from there. That might mean running research, architecture, implementation, and testing in parallel, then using your judgment to bring the results together.

You keep your own queues filled. While one line of work is blocked or running, you can prepare the next batch, test an assumption, improve the context, or advance another part of the mission. You know when to ask for a decision, but ordinary ambiguity does not stop you.

This does not mean generating the maximum amount of work. It means building the fastest reliable loop from mission to evidence to a product we would stand behind.

Who we're looking for

An engineer with excellent fundamentals, strong product judgment, and enough range to work across the stack. You do not need to be the strongest specialist in every layer, but you should be able to learn unfamiliar territory quickly, make good tradeoffs, and verify work you did not write line by line.

You already use AI heavily, and your workflow has moved beyond asking one agent to perform one task while you wait. You know how to decompose work, run useful tasks in parallel, preserve context, review generated work, and improve the system when the same bottleneck appears twice. If you are earlier in your career, we care about the trajectory: evidence that you are actively increasing what you can own, and that your fundamentals hold up under review.

You care about how the product feels as much as whether it functions. Complexity is information for the plan, not a reason to avoid the attempt.

Technical areas that matter

Useful experience includes some combination of:

• Python and backend application development
• data modeling, event pipelines, experimentation systems, or metric integrity
• LLM application architecture and multi-agent engineering workflows
• SDK design, integrations, or developer experience
• production systems with clear verification, observability, and rollback
• TypeScript and modern frontend application architecture
• browser automation, visual regression testing, and product analytics

You do not need all of these. We care about a strong foundation, visible product judgment, and evidence that you can use agents to cross boundaries without lowering the quality bar.

Strong signals

Strong evidence might be a product you took from a rough goal to a shipped result, a system whose verification you clearly cared about, or an AI-native workflow where you can show how parallel work was created, checked, and integrated. We are interested in the system you built around yourself as much as the final artifact.

The exact background matters less than whether you can take a higher-order mission and turn it into a working, trustworthy product without waiting for someone else to produce every task.

Location

San Francisco is strongly preferred. We want to work together in person and build shared context quickly during this stage of the company. We will consider an exceptional remote engineer with strong evidence of independent, high-throughput work and excellent written communication.

How we work

We run the company from shared, structured context that people and agents both use. We build with spec-driven development, flights of agent work, automated verification, and human review at the decisions that matter. Every engineer is expected to improve that system alongside the product work.