What You’ll Build:
- Social AI Agent: You know those people in your office, in your friend group, or that friend of a friend: the ones who know everyone, the ones who make things happen, introduce people to each other, and invariably host things at their house.
Think of our agent as that friend, for everyone:
- It knows every hidden gem in your city: events, venues, people, groups.
- It knows your whole crew’s vibe, taste, and availability.
- It has the agency to go from idea -> plan -> booked -> reminded -> done.
- It introduces you to people who'd become real friends.
It gets better continuously, in three ways:
- Memory: Knows you better with every conversation.
- City Intelligence: Inputs from all users feed everyone’s recommendations, both for things to do and for people.
- Network: More users = better plans, better people, better matches.
You'll build this agent with the CTO:
- Build production agents from scratch: prompts, system design, workflows, tool integrations, and evals.
- Map the space the agent has to cover. Work out the scenarios that matter, how each should be handled, and where the edges are, so coverage is designed rather than discovered later.
- Optimize agent performance through prompt engineering, model selection, tool design, skill writing, context window management, and eval harness development.
- Own quality end to end. Define what good means for the use case, build the evals that measure it, and close the gap until the numbers hold.
- Engineer agents like software. Version them, review them, and hold them to a regression suite, so quality does not depend on who last edited a prompt.
- Improve agents' ability to plan, use tools, manage context, recover from errors, and complete long-running tasks reliably.
- Build secure, observable, and reliable agent systems, including permissions and safeguards for sensitive actions. Develop tracing, replay, and monitoring infrastructure that makes agent failures reproducible and actionable.
- Analyze real-world failures and use data to drive iterative improvements.
- Reliability, latency, and cost efficiency are product features here, and you'll own them in production.
Software Factory:
Our mantra: LLMs are not here to help us write better code faster. Our job is to help the LLM write better code faster. AI needs to be able to work autonomously and continuously with less and less human babysitting. By \"software factory, \" we mean work that runs on its own. It starts from a trigger, whether that's our data, logs, a bug report from a user, or a new feature requirement from the team, and it ends at shipped things that work. In between, agents write the code, babysit the PR until it's green, review it, and handhold the change until it is safely and fully rolled out into production.
Your job is to:
- Extend agentic workflows beyond initial code draft generation into test creation, automated code review, agentic deploy and monitoring, CI failure triage, and release verification.
- Identify problems that the LLM cannot solve on its own, and help it not only solve them this time but also solve them better, autonomously, next time.
- Build agents that take on repetitive engineering work, such as diagnosing failures, fixing broken tests and latency regressions, and running migrations.
- Define what it means to trust a change, especially changes written by AI, and build the tooling that holds the bar throughout the software development lifecycle.
- Decide what agents are scoped to do in the pipeline, what tools and permissions they hold, how they are sandboxed, and what they are not trusted to do unsupervised.
Infrastructure and Reliability:
- Own the reliability, scale, performance, and cost of our systems.
- Own environment strategy, secrets handling, deploy observability, and kill-switch and rollback mechanics.
- Build agents that sift through alerts and dashboards, de-duplicate signals, identify real latency regressions, root-cause them, and propose fixes.
Your Skills and Experience:
- We care far more about what you have built and how you think than about college, brand names, or years of experience.
- You have shipped something to production and owned it afterwards. Not just built it. Kept it running, fixed it at 11pm, and let real usage change your assumptions.
- You have a live AI product or a GitHub project you've built that you can walk us through.
- You have worked seriously with LLMs and agents: context engineering, tool use, structured outputs, memory, evals, and the failure modes that only show up at scale.
- You have experience improving AI quality through evals, prompt/system design, retrieval quality, feedback loops, and human review.
- You bring engineering discipline to AI. You think in versions, tests, regressions, and reproducibility. You are uncomfortable when a system's behavior depends on a prompt nobody can account for.
- You have strong Go and range across the stack: backend services, data pipelines, APIs, cloud infrastructure. When you meet an unfamiliar stack, you pick it up in hours.
- You're proficient with cloud infrastructure (AWS) and infrastructure as code.
- You have a good understanding of PostgreSQL and relational databases.
- You have good product sense: you can decide what not to build and simplify a messy problem.
- You're excited by 0 -> 1 work and comfortable navigating ambiguity.
- You turn ambiguous business and user problems into simple, usable product flows.
- You make pragmatic architecture choices and ship frequently.
- You have an iterative mindset: you develop proofs of concept, make decisions quickly, and ship v0s.
Even Better If You:
- Are passionate about social apps and believe the consumer space is broken and ripe for reinvention.
- Have been a founder or founding engineer and know what it means to balance craft, ownership, and speed.
- Have experience at a startup or on a similarly small team, taking a product from zero to one and talking directly to your users.
- Have a point of view on how software engineering changes as agents write more of the code.
- Have taken an early consumer product to something people use and love.