Kaavio builds AI-powered tools that help people buy and sell complex physical products. We work with industrial and science distributors managing massive catalogs (millions of SKUs, each with detailed specs and docs), extracting images, descriptions, and structured data for 10k+ products in days instead of months.
The opportunity is enormous: B2B buyers will spend $8.5T online by 2030 (5x consumer ecommerce, growing twice as fast) yet 83% of B2B sellers say their product data is incomplete or outdated. This role sits at the heart of what makes Kaavio work: turning messy product data into trustworthy content at scale.
We were founded by a CEO with prior founder experience and deep expertise in technical sales, alongside a CTO who has spent years building and leading engineering teams at startups. Learn more about us here.
- Start with the customer: we build for real problems, learned from customer conversations.
- Learn by doing, not debating: write code, ship something, or talk to customers instead of debating hypotheticals.
- Work like scientists: especially in AI, the only way to know we're making progress is to run experiments and measure results. The difference between screwing around and science is writing it down.
- Everybody codes & everybody talks to customers: anyone can push a CSS fix or write a SQL query, and everyone gets on customer calls.
- Default to sharing: we share everything openly: draft PRs, planning docs, company plans, board notes, and finances.
- Celebrate wins: big and small, including the mistakes we learned from.
- Problem owners who take a problem and make it theirs all the way through. Progress matters more than perfection.
- Resilient problem solvers who collaborate, iterate, and keep moving forward through startup messiness.
- Collaborative professionals who are thoughtful, flexible, and respectful.
- Genuine builders who love making things, software or not.
- Scientists at heart who reach for a measurement before an opinion and let the results change their mind.
- Building the AI systems behind our product: research agents, RAG over large catalogs, structured extraction from messy sources (PDFs, spec sheets, safety data sheets), and entity resolution.
- Treating AI as a measurement problem: we run curated eval sets on every PR and trace every agent call. You'll build evals and golden datasets, define metrics, run experiments, and ensure changes make the product better, not just different.
- Pushing past brute-force LLM calls: the naive approach (an LLM call for every problem) gets slow, costly, and inconsistent fast. Some of our most interesting work is what comes next, at the intersection of data science and engineering.
- Working with data at scale: building pipelines that ingest, normalize, and enrich catalogs of tens of thousands to millions of SKUs, with care for correctness, cost, and throughput.
- Owning features end-to-end: shaping the idea, building an MVP, testing in production, and iterating — including thoughtful full-stack UI, pulling in frontend and design support as needed.
- Talking with and supporting customers: joining customer calls to ask questions and share early work, plus rotating into support to solve issues (some on-call time is part of the role).
- Strong software engineering fundamentals: you take a system from idea to deployed and stable.
- A data science or applied-ML background: comfortable with experimentation, evaluation, and reasoning about data.
- Experience building information retrieval, search, recommendation, or extraction systems (embeddings, ranking, RAG, or similar).
- Comfort with backend and data engineering: APIs, databases, and data pipelines (TypeScript, Postgres, and BigQuery are our stack).
- Comfort operating in ambiguity and taking initiative without perfect specs.
- Strong writing skills: you should be able to communicate clearly, concisely, and persuasively in writing.
- Must be legally authorized to work in the US. We are unable to provide sponsorship for this role.
- Shipped LLM-powered features in production, including the eval and observability work that makes them trustworthy.
- Experience with big data tooling (BigQuery, warehouses, large-scale ETL).
- A graduate background or research experience in ML, IR, NLP, or a quantitative field, paired with a track record of shipping.
- Experience with code-generation pipelines, classical ML, or turning probabilistic systems into deterministic ones.
- Experience at an early-stage SaaS startup or working closely with customers in a technical context.
- Salary: $180k – $220k / Equity: 0.25% – 0.75% / (depending on experience and location)
- Flexible working hours and location (US-based)
- Health, dental, and vision coverage
- Generous vacation policy – take time when you need it
- A direct path to shaping technical direction, mentoring, and leadership as the team scales
- Direct customer exposure and high impact on product direction
We aim to be efficient and transparent. You'll be Kaavio full-time employee number 4. Our process centers on spending a day working together in-person on actual projects. We'll fly to you or you to us and grab a co-working space for the day.
- Intro call – 30 min with a founder to understand your motivations and answer your questions.
- Technical conversation – meet an engineer to talk through a past project and a short systems/design exercise, no live coding.
- Work with us – join us for some real work and ship code.
- Final conversation – we'll answer any remaining questions.
- Offer – 🎉
If you have a disability, please let us know how we can make the interview process better for you — we're happy to accommodate!
Interested? Send us a note at founders@kaavio.ai with the information:
- Your name
- Where you’re located
- A few sentences about why you’ll be a great fit for us and this role
- Your resume or CV
If this sounds exciting, we’d love to hear from you
