Senior Engineer, AI & Data

Clausis · Budapest · Hybrid · Full-time

About Clausis:

Clausis is an AI platform for legal professionals in Central and Eastern Europe. Lawyers use it to research legislation and case law, work with their documents and draft faster, in their own language and under their own jurisdiction. Hungary is our home market and we are expanding across the region. We are a small team with paying customers, and we are building the next generation of the product now.

What you will do:

Product

• Own large parts of a TypeScript codebase end to end: a Next.js/React web app, Node services, background jobs, and our mobile and Office clients.
• Build LLM features that hold up in production: tool calling, streaming, retrieval, multi-step agent workflows, and the evals that show whether a change made answers better or worse.
• Keep answers fast and costs in check, including model routing across several providers.

Data

• Turn messy HTML, PDFs and scans into clean records: parsing, OCR, metadata and date extraction, deduplication, chunking, embeddings and hybrid search.
• Keep it running unattended: nightly incremental updates, monitoring, and catching the source that silently stopped updating.
• Improve retrieval quality and the cost of storing and searching millions of documents.

How the work gets done

• AI coding agents write most of our code. You write their briefs, review their pull requests hard, and set the guardrails that keep their output shippable.
• Talk to the lawyers who use the product, see what fails, and fix what matters most.

You have:

• 5+ years building production software in TypeScript, Node, React/Next.js.
• Production experience with LLMs: tool calling, RAG, evals, and debugging bad model output.
• Data pipelines or scrapers you built and ran unattended, and you know how they break: layout changes, rate limits, broken PDFs, encodings, duplicates.
• Security sense for multi-tenant SaaS: auth, tenancy, permissions, audit trails.
• Judgement to ship small, safe changes in a fast-moving codebase, and to say no to scope.
• Clear written English. Hungarian is a strong plus; another CEE language is a plus.

Nice to have:

• Python for data work: PDF and OCR tooling, data analysis, NLP or ML experiments.
• Embeddings and vector search (pgvector, Pinecone or similar), ranking and relevance tuning.
• Playwright or Puppeteer, Tesseract, PDF and DOCX internals.
• Vercel AI SDK, MCP, Drizzle, Expo.
• Legal tech, fintech or another regulated domain; early-stage startup experience.

How we work:

• We move fast and ship daily. This is not a 9-to-5 job: around launches and customer pilots we work evenings and weekends, sometimes around the clock. If you want a steady, predictable week, this role is not for you.
• Your leverage comes from directing and reviewing work, not just typing it.
• Decisions are written down; work ships in reviewable pull requests and is checked live.
• What you build reaches paying customers within days.

We offer:

• Competitive salary, based on your qualifications and experience.
• Hybrid: Mostly remote some days at our Budapest office
• Flexible hours.