AI Engineer - LLM Inference Stack

About the role

An AI Engineer is needed in Islamabad, working onsite, to turn research into reliable, customer-facing products. You will architect and own the inference stack behind LLM-based APIs, RAG pipelines and autonomous agents, then scale it to thousands of concurrent users.

Key responsibilities

• Build and harden the core API layer exposing large language models to internal and external clients, covering latency, authentication, rate limiting and multi-tenancy.
• Design and ship robust RAG systems (chunking, embedding, retrieval, re-ranking) and agentic workflows (planning, tool calling, reflection loops).
• Architect end-to-end ML pipelines that scale horizontally on AWS or GCP while keeping cost per token predictable.
• Evaluate new model families, fine-tuning techniques and orchestration frameworks, running A/B tests to measure real-world gains.
• Optimize prompts, inference kernels and serving stacks for throughput, latency and memory footprint.
• Write production-grade Python, extensive unit and integration tests and clear design documents.
• Work daily with product, design and backend teams in a fast-moving environment.

Required profile

• Hands-on experience with at least one LLM ecosystem: OpenAI APIs, LangChain, LlamaIndex, Vertex AI or open-source models.
• Strong grasp of software engineering fundamentals: version control, testing, CI/CD and observability.

Nice to have

• Experience with vector databases (Pinecone, Weaviate, Milvus) and embedding models.
• Familiarity with inference optimization tools such as vLLM, TensorRT-LLM, ONNX and quantization.