Company: TaxDown
Country: Spain
Type: Onsite
Employment: Part-time
Description:
About the role Build the AI operating system of TaxDown's tax team: transform manual processes and expert knowledge into intelligent, measurable and scalable flows. You will take a domain of the tax team (case review, documents, procedures, support) and make it yours from end to end: understand it, redesign it, build the solution and answer if it has worked. Why this problem is interesting At TaxDown we help people free themselves from bureaucracy. Behind that experience is a team of tax experts that resolves doubts, reviews cases, fills out models and executes procedures. This puts you in front of the problems that are really difficult to automate: unstructured documents, thousands of conversations with users, decisions that require context, and processes where precision really matters. You will build internal products that multiply the capacity of our tax experts. You will work directly with tax experts, responsible for Operations, Data, Product and Engineering, with access to real problems and the ability to take your solutions to production. AI is already part of the team. We have systems in production working alongside our tax experts on problems that almost no one has solved well yet. The commitment to AI is one of the company's annual objectives. What you will do Applied AI systems: Systems and data engineering: - Robust integrations through APIs, webhooks, queues and events. Process design: What we are looking for We are looking for someone who enjoys building and who has already seen what happens when what you build is actually used: the rare cases, the silent errors, people who use it differently than you imagined. Engineering or quantitative training, Telecommunications, Computing, Industrial, Mathematics, Physics, Data Science or similar, or demonstrated ability to think like this. We look for that way of reasoning: systems, data flows, bottlenecks, traceability and measurement. Requirements: 1-2 years of experience building: automations, internal tools, analysis, product or engineering. A startup, a consulting firm, a data team or an operations team are worth it. Programming: Python or JavaScript, and SQL. You don't have to be a senior developer, but you do have to write code, understand a REST API, a webhook and a JSON without help. Structured thinking: take a fuzzy problem and turn it into a diagram, some steps and a metric. Autonomy and communication skills: knowing what to ask, to whom and when. Our culture Beyond experience, there are traits that the people of TaxDown share: honesty, judgment and a real desire to learn. We are looking for someone who takes responsibility, works with ambition and resilience, and enjoys helping and solving problems with impact. We want taxes to be simple and accessible to anyone. We would like to enhance people's decision-making with technology. Fight for every penny that rightfully belongs to them. Provide advice to ensure that your taxes are done correctly, and in good faith. We want to serve a simple solution that solves a global problem: doing your taxes well, optimizing your opportunities. We are looking for problem solvers, curious thinkers and creators of opportunities, but, above all, people who seek to create a real impact in the world. What we offer Indefinite contract. Real ownership and fast production. What you build is used the next day and frees up hours of real people. Direct exposure to good people: tax, Operations, Data, Product and Engineering experts, with access to real problems and the ability to take your solutions to production. A role with a journey. In one year you will have designed reliable AI systems, integrated tools via API, modeled complex processes and taken several solutions from idea to production. The Stack Python, JavaScript/TypeScript, SQL, OpenAI and Anthropic APIs, n8n, Airtable, Metabase and AI-assisted development platforms. You will choose the right tool for each problem: from a workflow in n8n to a service in Python, an integration via API or a complete internal application. We don't expect you to know all this. Yes, you have built with part of it and learn the rest quickly. It is valued: - Having put a system with LLMs in front of real users.
- Agents capable of executing multi-step operational tasks.
- Tax document extraction and classification systems.
- Internal assistants that combine tax knowledge, user context and operational tools.
- Experimentation with RAG, tool calling, structured outputs and multi-agent flows.
- Human in the loop mechanisms for experts to review, correct and improve the system.
- Data modeling and pipelines that support operational reporting.
- Management of errors, retries, permissions, logs and observability. What you build has to survive the rent campaign.
- Criterion to decide when to use deterministic rules, traditional automation or AI. And criteria for not building: sometimes the correct response is to change the process, not automate it.
Sit down with the tax team, understand how they work today and find where time and context are lost. Propose the "how it should be" and defend it with data. Measure if what you have built has improved something. If not, revert it.
- Experience in high operational volume environments: support, back office, customer ops.
- n8n, Make, Airtable, Retool, Replit, Softr, Notion.
- Metabase, dbt or experience setting up dashboards and modeling data.
- Intercom, HubSpot, Customer.io.
- Own projects or side projects. We are more interested in what you have built than where you have worked.
Apply here:
Web: Apply here
Emails:
Found 6 similar Onsite jobs
United States
View Job →