Staff AI Engineer

We're team Variant Group — we build B2C SaaS products that millions of people use every day: pdf.net, resume.co, resumetemplates.net & cv.co, and others.

We're fully remote, operating worldwide: USA, Spain, Germany, Serbia, Argentina, Thailand.

We're hiring a Staff AI Engineer to own the AI behind pdf.net — the agentic systems, document intelligence, and LLM features that power the product. You'll set the technical direction for the AI stack, own how reliably and predictably it ships, and be the person who makes sure AI features land on time and stay up at scale.

This is a hands-on, high-ownership role. You'll work across the full lifecycle — from architecture and agent design to latency, cost, reliability, and real user impact — and you'll be the connective tissue between the AI work and the rest of the company.

What you'll own

• Delivery. Ship AI features end to end and own that they land on time and predictably. Set a realistic pace for the team — and hit it.

• Reliability. Keep production healthy: monitoring, incident response, fallbacks, release safety. When the AI app misbehaves, you see it coming and fix it — not "well, that happened."

• Technical direction. Own architecture and build-vs-buy calls for the AI stack. Know when to adopt a framework and when not to; keep us out of tech debt we can't climb out of and away from dependencies baked so deep we can't move.

• Communication. Translate AI work into plain language for product and other leads — what's happening, why it's hard, and how long it'll really take.

• Cross-team work. Partner cleanly with backend, product, and design.

• Raising the bar. Lift the AI team through code review, mentoring, and solid engineering practice — and spread knowledge so the system isn't trapped in one person's head.

• Proactivity. Catch problems before they become incidents, and bring them forward yourself.

What you'll do

• Design and ship agentic, multimodal, evaluation-driven product features across extraction, reasoning, generation, and orchestration.

• Work on document intelligence problems: OCR, structured extraction, chunking, layout-aware parsing, and PDF-to-schema workflows.

• Design and ship RAG pipelines (LlamaIndex) using vector databases, hybrid retrieval, reranking, and structured context assembly.

• Integrate hosted LLM APIs (OpenAI, Google Gemini) and, where needed, smaller self-hosted models (served with vLLM) into product features.

• Build scalable inference services using batching, streaming, caching, and other latency/cost optimizations.

• Implement production pipelines for evaluation, prompt/version management, and deployment, with offline evals and regression datasets for accuracy, latency, and reliability.

• Own AI-related infrastructure in Docker, Cloud Run, and GCP, with proper logging, monitoring, and observability.

What You Bring

• 8+ years of engineering experience, with a track record of owning systems in production.

• Deep Python — you build solid services (FastAPI) with sensible tests and understand how it works under the hood. Not "I write scripts."

• Pragmatic delivery. Clean, enterprise-grade code matters — but shipping a working result in a reasonable timeframe matters more. You know when "good enough" ships, and you don't gold-plate.

• Proven experience building agentic / LLM systems in production — not just prototypes. You know the real-world traps: frameworks, memory, streaming, evals, cost, latency.

• Hands-on experience integrating OpenAI / Gemini APIs into real user-facing products.

• Practical experience with RAG systems: chunking, embedding pipelines, vector search, reranking, context construction, and hallucination reduction.

• Experience with vector stores (pgvector, Qdrant) and hybrid search; judgment on when retrieval helps and when a deterministic pipeline is better.

• Familiarity with document understanding workflows: OCR, PDF parsing, table/form extraction, normalization of messy real-world inputs.

• DevOps & MLOps at a real level — containers, GCP, CI/CD; plus the MLOps side: model serving, eval pipelines, and LLM tracing/observability. You own your systems end to end instead of paging infra for every change.

• Strong communication and collaboration, comfortable working remote/async, self-motivated.

Working with us

• 🌏 100% remote, work from anywhere
  Live wherever you want. We just need some overlap with Asian and European time zones for coordination.

• ✈️ Regular meetups and offsites
  We meet in person regularly — hackathons, workshops, team trips. The company generously covers flights and accommodation.

• ⚡️ No micromanagement
  We measure results, not hours on Zoom. Do great work and enjoy real autonomy.

• 🏖 Unlimited time off
  We trust you to decide when you need to rest and recharge.

• 🧠 Work with professionals
  The team includes talented specialists in engineering, design, product, and operations. We experiment, learn, and make decisions based on data.

• 📚 Growth investment
  The company pays for courses, software, subscriptions, and conferences. Want to level up a skill? We've got you covered.

• 🤖 Working on the AI frontier
  We don't just build AI products — we use AI in our daily development. Claude Code, Cursor, Codex, various integrations and MCPs, five AI reviewers on every PR. Any tool that makes you more productive, the company will pay for, no questions asked.

• ❤️ Wellbeing
  Health & Wellness Stipend (e.g., gym, fitness, general wellbeing expenses).
  Mental Health Support (including access to an in-house mental health consultant).

• 💰 Salary & Contract
  $90,000–$132,000/year gross, paid via deel.com

How to apply?

1. Submit CV.

2. We'll invite you for a 1-hour screening call.

3. We'll invite you for a 1-hour technical interview.

4. If we see a potential match, we might ask you to complete a small challenge that requires up to 8 hours of your time.

5. Receive the offer 🎉