Product · Budapest · Hybrid · Full-time
Build software that
has to get the answer right.
Most software can survive being slightly wrong. Professional tax and legal work cannot.
azta is building an AI co-worker that researches the law, works through professional documents, follows legal references, drafts deliverables and shows the source behind every conclusion. Across 24 European jurisdictions, our job is to turn complicated professional work into a product someone can trust enough to use with a real client.
We’re looking for a Product Engineer who wants the whole problem — not just the frontend ticket or the backend endpoint. You’ll sit where product, engineering and AI meet, and take customer problems all the way from a rough idea to a production workflow.
Team
Product
Location
Budapest, 13th district
Working setup
3 days office · 2 days home
Type
Full-time
What you’ll work on
From professional workflow to product
Core product
You’ll build the surfaces professionals use every day:
- research and chat workflows,
- statute exploration,
- cited answers,
- document upload and analysis,
- source viewers,
- document generation,
- Canvas workflows,
- editing experiences,
- collaboration and sharing,
- and professional workflows we haven’t built yet.
azta already turns research into editable summaries, advisory letters, emails and Word documents rather than stopping at the answer. Your job is to make all of that feel like one coherent product.
AI as product infrastructure
You do not need to be an ML researcher. You do need to be comfortable building software where models and agents are part of the system.
You may work on:
- agent tool interfaces,
- streamed research states,
- model output represented safely in the UI,
- citations and verification,
- structured outputs,
- human confirmation flows,
- retrieval experiences,
- prompt-driven product behaviour,
- evaluation,
- and graceful failure when the model cannot do what the user asked.
In 17 of the 24 European jurisdictions we cover, the structured legal engine walks statute hierarchies and resolves cross-references rather than relying entirely on flat retrieval. That creates product problems most SaaS engineers never get to touch.
Documents
A professional rarely starts with a blank chat box. They arrive with a contract, a statement, a spreadsheet, an email, a memo or a previous piece of advice.
You’ll help build the systems that:
- ingest those files,
- surface the relevant evidence,
- connect a citation back to the passage it came from,
- let users edit the generated work,
- and turn the result into something they can actually send.
Enterprise
Some azta customers do not use the shared environment at all. We run dedicated deployments with their own domain, database, authentication and configuration.
You may work on:
- organisation-level configuration,
- permissions,
- SSO,
- knowledge space experiences,
- a firm’s own internal knowledge,
- branded outputs,
- workspaces,
- APIs,
- and features that have to behave differently for a hundred-person firm than for an individual Pro user.
Product discovery
Engineers here are expected to understand why something should exist.
You’ll:
- sit in customer conversations,
- watch professionals work,
- prototype ideas,
- challenge requirements,
- ship the smallest useful version,
- and iterate from what people actually do with it.
“Product Engineer” is intentional.
Who we’re looking for
You care about the last 10%
You may be a great fit if you:
- Are an excellent TypeScript and JavaScript engineer.
- Have strong React and modern frontend experience.
- Work comfortably across APIs, databases and backend systems.
- Know SQL well enough to investigate the product yourself.
- Have strong product judgement.
- Care how the workflow feels, not only whether the endpoint returns 200.
- Can navigate an unfamiliar codebase independently.
- Like ambiguous problems.
- Are happy talking to users.
- Ship quickly without treating quality as optional.
- Think hard about edge cases, especially when the output carries professional consequences.
It’s a strong plus if you’ve worked with:
- Next.js
- PostgreSQL
- Supabase
- AI agents
- LLM applications
- retrieval systems
- rich-text editors
- document processing
- OCR
- search
- enterprise SaaS
- regulated products
- Python
We do not expect every engineer to arrive knowing European tax law. We do expect you to be curious enough to learn why the details matter.
How we work
Engineers stay close to the work
You’ll work directly with Product, AI Engineering, Design, our domain experts and our customers. A lawyer may explain a workflow in the morning and you may have the first version in production days later.
There are few handoffs and little room to hide behind a ticket. That also means real autonomy: if you see a better way to solve the problem, you’re expected to say so and build it.
This is a hybrid position with regular in-person collaboration at our Budapest office (13th district, Váci Road): 3 days from the office, 2 days home office per week.
What we offer
The whole product, not one layer
- End-to-end ownership of meaningful product surfaces.
- Direct access to customers and domain specialists.
- Hard technical problems across AI, search, documents and structured law.
- A product used for real professional work.
- A small, experienced engineering team.
- Direct collaboration with the founders.
- Fast release cycles.
- Competitive compensation.
- Real equity.
- Hybrid working with 2 home-office days per week.
You’ll see your work reach customers quickly — and you’ll know exactly why you built it.
Apply
What have you shipped?
Send us your CV, LinkedIn or GitHub and point us to something you built where you owned more than the code. We’d especially like to hear what the original problem was, which decisions were yours, and what changed once real users touched it.
By applying, you consent to the processing of your data as described in our Candidate Privacy Notice. Prefer email? Write to jobs@azta.ai.