Roles & Responsibilities :
This role architects production-ready AI and GenAI solutions that are reliable, observable, safe, and cost-efficient. The architect owns the technical design of AI solutions from opportunity through delivery. The solution architect is the senior technical authority on AI architecture decisions within projects and collaborates with the Data Solution Architect where engagements span both data foundations and AI solutions.
Key responsibilities
AI & GenAI solution architecture
Own end-to-end architecture for AI and GenAI engagements — classical ML, LLM applications, RAG systems, and agentic workflows.
Responsible for design decisions: build vs. buy, model selection, fine-tuning vs. retrieval, and orchestration patterns.
Design retrieval architectures (chunking, embedding, indexing, hybrid search, and re-ranking) for accuracy and latency at scale.
Architect agentic systems: tool use, memory, multi-step orchestration, and human-in-the-loop control points.
Production readiness & LLMOps design
Design LLMOps/MLOps setup for engagements: model serving, versioning, deployment pipelines, and rollback.
Define evaluation architecture: eval harnesses, quality benchmarks, regression testing, and acceptance criteria.
Architect observability, monitoring, and drift detection; design for operability and handoff to Managed Ops.
Design cost-efficient inference architectures: model tiering, caching, token economics, and FinOps guardrails.
Data foundations for AI
Specify the data requirements for AI solutions — training data, feature pipelines etc.
Partner with Data Solution Architects to translate model and RAG requirements into data platform design.
Design embedding and vector store strategy in sync with the underlying data architecture.
Safety, governance & non-functional design
Embed Responsible AI and AI security standards into solution design — guardrails, explainability requirements, and human oversight.
Design against GenAI risk surfaces: prompt injection, data leakage, unsafe outputs, and insecure tool use.
Design for non-functional requirements: latency, scalability, availability, security, and cost. Ensure designs meet regulatory obligations.
Delivery & engagement support
Serve as technical authority through delivery — guiding engineering teams, reviewing designs, and resolving technical escalations.
Support pursuits with technical proposals, effort estimation, and technical workshops with client stakeholders.
Experience
10–15 years in software/ML engineering and architecture, with proven ownership of AI solutions in production
Expected Skills
Strong Hands-on GenAI architecture — LLMs, RAG, agentic patterns, orchestration frameworks and tools such as Langchain, Haystack, and Llama Index
Classical ML delivery grounding
MLOps/LLMOps, evaluation design, model serving, observability, and inference cost optimisation
Practical experience designing to Responsible AI, security, and compliance requirements
Working knowledge of data platforms, pipelines, and modelling.
Exposure to Enterprise Architecture is an added advantage
Ability to lead technical workshops with technical stakeholders in client environment
Excellent verbal and written communication, technical authoring, ability to communicate technical concepts and trade-offs to stakeholders of varying technical competency
The boar re-runs this search every day and emails you what’s new — scraped straight from company career sites, no recruiter spam. Free.