Principal AI Architect

AI Architecture & System Design
• AI system architectures: multi-agent orchestration layers, RAG pipelines, hybrid retrieval systems (knowledge graphs + vector search), text-to-SQL engines, and real-time inference APIs.
• Define and own technical blueprints for new AI products - from data ingestion and embedding pipelines through to response generation, evaluation, and production monitoring.
• Solve hard engineering problems: latency, precision/recall trade-offs, context window management, hallucination mitigation, and cost-efficient LLM usage at scale.
• Make deliberate, well-documented architecture decisions with clear trade-off analysis (build vs. buy, framework selection, deployment topology).


Implementation
• Write production-quality code - Python, SQL, API services - across the full AI lifecycle: data qualification, model training, evaluation, containerised deployment, and API serving.
• Build and own reusable, framework-quality components (chunking pipelines, retrieval layers, agent tool-calling modules) that accelerate team velocity.
• Own CI/CD pipelines, Docker-based deployment, and production telemetry for AI services.


AI Market Intelligence & Technology Strategy
• Track and evaluate the AI landscape - new LLMs, agentic frameworks (LangGraph, Google ADK, CrewAI, AutoGen), retrieval methods, fine-tuning techniques, and emerging tooling.
• Translate AI market trends into actionable roadmap inputs - surfacing opportunities for step change capability improvements before competitors do.


Cross-Functional Technical Partnership
• Partner closely with Product, Data Science, and Platform Engineering to align AI architecture with product direction, data constraints, and infrastructure capabilities.
• Communicate complex technical trade-offs clearly to non-technical stakeholders - translating architecture decisions into business impact narratives.


Must-Have Experience
• 12+ years of hands-on experience in AI/ML engineering and data science, with significant depth in production system delivery.
• Deep, working expertise in LLM application development: LangChain, LangGraph, tool-calling agents, RAG, prompt engineering, embedding pipelines, and hybrid retrieval.
• Proven track record architecting and shipping multi-agent systems, knowledge graph-powered retrieval (Neo4j or equivalent), and real-time inference APIs.
• Strong ML fundamentals: XGBoost, deep learning, NLP, time-series forecasting, propensity modelling, experimental design, and causal inference.
• Experience delivering AI systems in regulated industries (financial services, cybersecurity, healthcare) with SOX, GDPR, or SOC 2 compliance awareness.
• Expert-level Python and SQL; fluency with GCP, AWS, Docker, FastAPI, BigQuery, FAISS, and CI/CD tooling.


Technical Depth
• Ability to design hybrid retrieval architectures that balance precision (graph traversal) and semantic recall (vector similarity), with reranking layers - not just off-the-shelf RAG.
• Hands-on experience reducing LLM inference latency in production (e.g., redesigning pipelines from multi-minute to sub-30-second response times).


QUALIFICATIONS
• Master's or PhD in Computer Science, Operations Research, Statistics, or a related quantitative field
• AWS Certified Machine Learning Engineer or GCP Professional ML Engineer certification.
• Completion of an AI Strategy or AI Governance programme.
• Prior experience at a data science / ML services firm, enterprise SaaS, or fintech - where you shipped AI to external customers, not just internal tools.
• Hands-on experience with Snowflake Cortex or comparable enterprise LLM deployment platforms.
• Open-source contributions to AI/ML tooling, published technical writing, or conference presentations.