AI Builder

What you will build

● Deal sourcing: systems that discover companies, founders and emerging signals before

they become obvious. Think agents that search, enrich, connect and rank fragmented

information - with a human investor firmly in the loop.

● Talent Radar: a living map of exceptional builders. Help us follow technical talent, founder

trajectories, team formation and the networks where the next company may emerge.

● Market intelligence: tools that turn messy public and private signals into evolving market

maps, company landscapes, theses and questions worth pursuing.

● The firm's memory: retrieval, research and knowledge systems that make every

conversation, memo and decision more useful the next time around.

● Precise rails: the less glamorous foundations that make everything else trustworthy -

data pipelines, identity resolution, permissions, evaluation, observability, feedback loops

and reliable cloud infrastructure.

Some of these products will look like agents. Some will look like copilots, workflows or quiet

infrastructure. You will choose the form that best solves the problem, with a bias toward

AI-native solutions over traditional software.

How you will work

● Gather requirements from all the team members and make a product roadmap

● Prototype in days, put work in front of real users, and learn from usage.

● Build evals and feedback loops alongside the product, not after it.

● Use human judgement where it creates edge; automate the rest.

● Make speed and precision allies. A sourcing system that is fast but noisy is not useful.

● Move comfortably from model behaviour and prompt design to APIs, data and production

infrastructure.

Your toolkit

We care more about what you can build than a checklist of credentials. You will likely be strong

in:

● Python and modern AI application development.

● Harness engineering: giving models the context, tools, memory, evaluation and

guardrails they need to do dependable work.

● LLMs, retrieval, agents, structured and unstructured data, and the judgement to know

when not to use any of them.

● Cloud systems and the production basics needed to make prototypes reliable, secure

and observable.

● Product sense: turning an ambiguous operating problem into a simple experience people

actually use.

● Clear communication with investors, founders and technical teams.

Who you are

● You are a builder first. A new model release gives you energy.

● You have shipped AI-native products or systems that people relied on, not only demos.

● You can hold a sharp product opinion and still update quickly when evidence changes.

● You enjoy messy, high-context problems where the answer is not already in a product

spec.

● You care about taste, speed and technical depth in equal measure.

● You want your work close to decisions, founders and the formation of new companies.