AI Architect - JB-2049

As an AI Architect, you will be responsible for architecting, designing, and deploying robust AI solutions with a strong emphasis on cloud-native architectures, machine learning pipelines, and generative AI technologies. You will collaborate with cross-functional teams to translate business requirements into technical blueprints, ensuring high performance, security, and scalability.

This role requires deep expertise in AWS ecosystems, advanced ML techniques, and the full spectrum of generative AI infrastructure.

Key Responsibilities

• Design and architect end-to-end AI systems, including data ingestion, processing, model training, inference, and deployment pipelines.
• Develop and maintain ML workflows encompassing supervised learning (e.g., classification, regression) and unsupervised learning (e.g., clustering, anomaly detection), incorporating tools like Optuna for hyperparameter optimization and XGBoost for gradient boosting models.
• Architect generative AI (GenAI) infrastructures, from foundational components like vector databases (e.g., integration with Amazon OpenSearch Service or third-party solutions like Pinecone for embedding storage and similarity search) to advanced agent-based systems. This includes designing Retrieval-Augmented Generation (RAG) pipelines, fine-tuning large language models (LLMs) using frameworks like Hugging Face Transformers, and building multi-agent architectures with tools such as LangChain or AutoGen for orchestration of autonomous agents.
• Ensure seamless handling of GenAI-specific challenges, including prompt engineering, token management, ethical AI considerations (e.g., bias mitigation in diffusion models or GANs), and infrastructure for real-time inference.
• Oversee the deployment of AI agents, including knowledge graphs, tool integrations (e.g., API calls, external services), and reinforcement learning from human feedback (RLHF) to create intelligent, adaptive systems capable of multi-step reasoning and decision-making.
• Work closely with DevOps teams to implement CI/CD pipelines for AI models using AWS CodePipeline, and monitor systems with Amazon CloudWatch.

Required Qualifications

• Bachelor's or Master's degree in Computer Science, Data Science, Engineering, or a related field; PhD preferred.
• 7+ years of experience in AI/ML architecture, with at least 3 years in a senior or lead role.
• Expert-level proficiency in AWS cloud services, including comprehensive knowledge of serverless architectures (AWS Lambda, API Gateway), containerized environments (Amazon ECS with Fargate), and database solutions (Amazon Aurora for MySQL/PostgreSQL-compatible relational databases, DynamoDB for NoSQL). Familiarity with AWS security best practices (IAM, VPC, KMS) and cost optimization tools is essential.
• Strong expertise in machine learning paradigms: supervised learning, unsupervised learning, and ensemble methods, with hands-on experience using Optuna for automated hyperparameter tuning and XGBoost for high-performance modeling.
• Deep understanding of the generative AI ecosystem, including:

--Vector databases and embeddings: Designing and implementing systems for storing and querying high-dimensional vectors (e.g., using Feature Store or integrating with Milvus/Pinecone), supporting applications in semantic search, recommendation engines, and multimodal AI.

--Large language models and transformers: Architecting solutions for model training/fine-tuning, handling datasets like Common Crawl or custom corpora, and optimizing for inference latency.

--Retrieval-Augmented Generation (RAG): Building hybrid systems that combine LLMs with external knowledge bases, including chunking strategies, embedding models (e.g., Sentence Transformers), and reranking mechanisms.

--Agentic AI infrastructures: Developing frameworks for AI agents, including multi-agent collaboration (e.g., using CrewAI or LangGraph), tool-use integration (e.g., function calling in OpenAI APIs), memory management (short-term/long-term), and deployment on scalable platforms like AWS EKS for Kubernetes-based orchestration.

--Advanced GenAI techniques: Experience with diffusion models (e.g., Stable Diffusion), variational autoencoders (VAEs), generative adversarial networks (GANs), and emerging areas like neuro-symbolic AI or federated learning for privacy-preserving GenAI.