AI Engineer – Agentic AI & RAG

Responsibilities
Design and develop agentic systems, including planners, tool orchestration, multi-step workflows, and guardrails.
Implement and optimise RAG pipelines, including indexing, chunking strategies, embeddings, hybrid search, reranking, and caching.
Build and deploy production-grade LLM applications, including prompt engineering, structured outputs, tool/function calling, memory, and context management.
Develop evaluation and monitoring frameworks, including offline test sets, automated evaluations, regression testing, and latency/cost tracking.
Enhance system safety and robustness through prompt injection defences, data redaction, policy enforcement, and access controls.
Own model deployment, scaling, API design, rate limiting, fallback strategies, observability, and incident response.
Collaborate with stakeholders to translate business needs into measurable ML outcomes, milestones, and delivery plans.
Mentor team members and contribute to engineering standards, code reviews, and best practices.
Qualifications
Minimum 2 years of experience in software or machine learning engineering with a proven record of deploying ML systems to production.
Engineering degree in Computer Science, Electronics, Machine Learning, or a related field.
Strong hands-on experience building LLM-based products including agents, RAG systems, tool calling, and structured generation.
Proficiency in Python and solid software engineering practices including clean architecture, testing, and CI/CD.
Experience with vector databases and search technologies such as pgvector, Pinecone, Weaviate, FAISS, Elasticsearch, or OpenSearch.
Understanding of distributed systems and production constraints including latency, cost, reliability, and security.
Practical knowledge of LLM evaluation methodologies and regression prevention.
Strong communication skills with the ability to lead workstreams end-to-end.
Nice to Have
Experience with model fine-tuning or adaptation (SFT, LoRA/QLoRA) and inference frameworks such as vLLM, TGI, or Triton.
Familiarity with knowledge graphs and structured knowledge systems (RDF/OWL, Neo4j, entity linking).
Experience in document AI and multimodal extraction, including OCR, layout understanding, and form/table processing.
Exposure to compliance-heavy domains such as healthcare or finance with privacy and security considerations.
Experience with agent frameworks such as LangGraph, Autogen, and Semantic Kernel, and orchestration patterns.

Work arrangement
Yes