About This Job
This is an engineering-first role: most of your time goes into infrastructure, deployment, CI/CD and reliability for ML and LLM systems in production, with hands-on work on retrieval pipelines and model adaptation when the product needs it. We're looking for around five years of experience shipping production software, applied to AI systems, rather than a research-oriented ML scientist.
WHAT YOU'LL BE DOING
Build and maintain CI/CD pipelines (GitHub Actions) and deployment configs for Docker Swarm and Kubernetes.
Own infrastructure automation and DevOps practices across environments; monitor production services.
Build and optimise LLM-based systems: RAG pipelines, embeddings, vector databases (FAISS, Pinecone, Weaviate or similar), prompt tuning.
Adapt models when needed (LoRA/QLoRA, PyTorch/TensorFlow) and build the evaluation harnesses that show whether the result %u2014 and its cost and latency %u2014 actually improved.
Apply LLM-specific security practices: prompt injection risk, sensitive data in retrieval pipelines.
Write automated tests and testing frameworks that scale with the codebase; take on technical debt rather than avoid it.
Communicate clearly with engineers, data scientists, product managers and business stakeholders.
WHAT WE'RE LOOKING FOR
Around five years shipping production software; a degree in computer science or equivalent practical experience.
Strong CI/CD, Docker, Kubernetes and infrastructure automation.
Deep Python; working familiarity with transformer-based models.
Strong fundamentals %u2014 reads, debugs and maintains code you didn't write.
Comfortable using AI dev tools in your own workflow and rigorous about reviewing what they produce.
Excellent communication across technical and non-technical audiences.
HOW WE ASSESS
Judgment over recall: a practical task with AI tools allowed, then interviews on the trade-offs behind your decisions.
A NOTE ON OUR HIRING PROCESS FOR THIS ROLE
We are actively hiring for a Senior AI Infrastructure Engineer. As part of this process, we have partnered with Lahore University of Management Sciences (LUMS) on a research study exploring how AI-assisted tools can support fairer, more effective technical hiring, and how AI coursework can improve AI readiness. Your application may be included in this study, but your personal data such as name, email etc. will remain confidential.
The hiring process consists of:
Optional Anthropic courses: applicants may be invited to complete these, and choosing not to will have no impact on their progress in the process.
Technical assessment
Technical interviews
Offer and onboarding