AI · Full-time
AI Engineer
Take LLM and vision features from prototype to production: retrieval, evaluation, guardrails and the cost engineering that makes them viable. We care more about your production instincts than your paper count.
What you will do
- Design retrieval pipelines — chunking, embeddings, hybrid search, reranking
- Build evaluation sets and automated scoring that runs in CI
- Engineer prompts as versioned, reviewed artefacts
- Implement guardrails: refusals, PII handling, injection defences
- Own the cost model for AI features and drive it down
What we are looking for
- Shipped an LLM-backed feature that real users used
- Strong Python and comfort with the modern model APIs
- Understand why a RAG system returns wrong answers and how to diagnose it
- Healthy scepticism about benchmark numbers
- Able to explain a trade-off to a non-technical stakeholder
Nice to have
- Multilingual NLP, particularly Indian languages
- Computer vision or on-device inference
- Speech recognition or synthesis experience
These are genuinely optional. We would rather hire someone strong on the core and curious about the rest than someone who ticks every box.
Hiring process
What happens after you apply
Application
Send us your CV or profile and a few honest lines about what you have built. We read every one and reply either way, usually within a week.
Intro conversation
Thirty minutes with someone who does the work you would be doing. Mutual — you should be interviewing us just as hard.
Practical exercise
A small, paid, take-home task close to real work, timeboxed to a few hours. No algorithm puzzles that neither of us will use again.
Deep dive
We walk through your exercise together, talk architecture and trade-offs, and discuss how you work with other people.
Offer
A clear written offer with compensation, scope and expectations. We do not do exploding deadlines or pressure tactics.
Ready to apply?
Send us what you have built and what you found hard about it. That tells us more than a CV.