About the role
We’re hiring a Prompt Engineer to own the LLM layer of our products.
Design prompts, build evaluation frameworks, architect RAG systems, and debug failure modes systematically. This is real engineering work — not “ChatGPT prompt influencer” stuff.
What you'll do
- Design, version, and maintain production prompts and prompt libraries
- Build evaluation frameworks and test suites that measure LLM output quality
- Architect and tune retrieval-augmented generation (RAG) systems
- Debug failure modes systematically — hallucination, drift, edge cases
- Partner with engineers to integrate LLM features into products
- Keep up with model releases and translate new capabilities into product value
What we're looking for
Must have
- 3–7 years of engineering experience, with real LLM/product work
- Hands-on experience designing prompts for production systems
- Experience building evals and measuring model output quality rigorously
- Practical RAG experience — chunking, retrieval, grounding
- Strong fundamentals in at least one of Python, TypeScript, or Java
- A systematic, experimental mindset — you debug with data, not vibes
Nice to have
- Experience with Claude, OpenAI, or other LLM provider APIs
- Familiarity with vector databases and embeddings
- Background in NLP, ML, or information retrieval
- Public writing or open source on LLM engineering
Why join AlpineQora
- Real ownership and autonomy — you drive your work, not just tickets
- Senior engineers who stay hands-on in the code, not just in meetings
- Hybrid work — 3 days office in Gandhinagar, 2 days WFH
- Comprehensive health insurance with family coverage
- Annual learning budget — courses, certifications, and conferences
- Performance bonus tied to real outcomes
- Work on systems for clients in fintech, healthcare, and other regulated industries
The process
A short, respectful process — typically 2–3 weeks from first conversation to offer:
- Intro call (30 min) — your background, what you’re looking for, and what we’re building
- Technical conversation (60–90 min) — a real problem in your domain, worked through together
- Portfolio / code walkthrough (60 min) — walk us through work you’ve shipped
- Final conversation — with the founding team
Notice period
Up to 60 days considered for strong candidates. Immediate joiners preferred.