EMIDS - Senior AI Architect

Job Description:

As a Senior AI Architect, you will lead the design and rollout of enterprise-grade Generative AI (GenAI) solutions, ensuring scalability, security, and business alignment. Beyond hands-on implementation, you'll play a pivotal role in defining AI strategy, architecture standards, and governance frameworks. You'll evaluate technology stacks, guide engineering teams, and collaborate with leadership stakeholders to bring real-time AI applications (summarization, conversational AI, decision-support systems) into production.

Key Responsibilities:

- Define end-to-end AI architecture blueprints, ensuring scalability, interoperability, and compliance.

- Evaluate and select appropriate LLMs, frameworks, and vector stores aligned with business needs.

- Lead design of Retrieval-Augmented Generation (RAG) pipelines and integrations with enterprise knowledge bases.

- Set architectural best practices for model deployment, monitoring, & optimization across cloud environments (AWS, Azure, GCP).

- Mentor AI engineers and foster cross-functional collaboration with product, data, and engineering teams.

- Stay abreast of GenAI trends (agentic systems, multimodal models, open-source LLMs) and advise on adoption roadmaps.

- Present architectural recommendations and influence AI strategy at the leadership level.

What We're Looking For:

- 12+ years in AI/ML, data engineering, or software architecture.

- Experience with agentic frameworks (AutoGen, Crew), MLOps pipelines, and security-conscious AI design.

- Track record of designing and deploying large-scale AI systems in production.

- Deep expertise in cloud-native AI deployments (Azure ML, SageMaker, Vertex AI).

- Hands-on knowledge of LangChain, Langraph, RAG, vector databases (Pinecone, Weaviate, Faiss).

- Strong background in Python and ML frameworks (PyTorch, TensorFlow).

- Proven experience guiding teams, making architecture decisions, and adopting technology.

- Strategic thinker comfortable with presenting to executives and technical stakeholders alike.