Artificial Intelligence Engineer

About the Company

A tier-one financial services enterprise is building next-generation AI infrastructure to power intelligent conversational systems—from chatbots handling routine transactions to voicebots managing complex workflows. They're scaling from proof-of-concept to production systems handling millions of interactions monthly.

About the Role

This is a senior engineering role focused on shipping robust, reliable AI applications on AWS, with direct impact on how millions of customers interact with financial services. You'll own the full stack: from designing RAG pipelines and prompt strategies, to orchestrating LLM workflows, to deploying containerized services that run 24/7. You'll work with leading models (including Claude via AWS Bedrock), define guardrails and responsible AI practices, and mentor mid-level engineers on production AI patterns.

Responsibilities

• Design and build production RAG systems — from chunking strategy and embedding model selection, to vector store architecture (pgvector, OpenSearch, Pinecone, or FAISS), to retrieval optimization for accuracy and latency
• Orchestrate complex LLM workflows using frameworks like LangChain, LlamaIndex, or Semantic Kernel; implement multi-step reasoning, tool use, and structured output patterns
• Develop conversational AI systems across text (chatbots) and speech (voicebots with STT/TTS integration); manage latency and streaming constraints in real-time voice
• Leverage AWS Bedrock and Claude models directly; design effective system prompts, few-shot examples, and chain-of-thought reasoning; iterate based on real-world performance
• Build and scale AWS infrastructure — Lambda functions, API Gateway, Step Functions, DynamoDB, SQS, S3; implement CI/CD pipelines with Docker/ECS or EKS; write infrastructure-as-code
• Define quality and safety guardrails — implement content filtering, responsible AI practices, and evaluation frameworks to catch model drift and output degradation before production impact
• Own API design (REST, GraphQL) and microservices architecture; build event-driven systems that integrate seamlessly with enterprise systems
• Mentor and unblock junior engineers; establish patterns and best practices for how your team ships AI features reliably

Qualifications

• 10+ years shipping production software in Python or Java (Python strongly preferred)
• 2+ years hands-on building and deploying AI/ML applications in production — not coursework, not personal projects; real systems in a real environment
• Deep RAG expertise — you've built chunking strategies, selected and tuned embedding models, chosen vector stores based on use case, and optimized retrieval for production
• Proficiency with AI orchestration frameworks (LangChain, LlamaIndex, Semantic Kernel, or CrewAI); you can explain the trade-offs between them and when to use each
• Hands-on AWS Bedrock and Claude experience — direct invocation, token counting, structured outputs, prompt optimization; familiarity with other LLM APIs (OpenAI, Azure) is transferable but AWS Bedrock depth is expected
• Advanced prompt engineering skills — system prompts, few-shot learning, chain-of-thought reasoning, tool use, and structured outputs aren't abstract concepts to you; you've debugged them in production
• Conversational AI systems — you've shipped text-based chatbots AND ideally have experience with voice systems (speech-to-text, text-to-speech, real-time streaming)
• Senior AWS proficiency — Lambda, API Gateway, Step Functions, DynamoDB, SQS, S3; you build without hand-holding and reason through cost and performance trade-offs
• CI/CD and containerization — solid grasp of Docker, ECS/EKS, infrastructure-as-code; shipping changes frequently and safely is table stakes
• API and systems design — you can articulate why you chose REST vs. GraphQL, how events flow through microservices, and where synchronous vs. asynchronous makes sense
• Evaluation and safety mindset — you've built evaluation frameworks for LLM outputs, implemented guardrails, and thought deeply about what can go wrong at scale

Preferred Skills

• Experience with model fine-tuning or retrieval optimization (RAG vs. in-context learning trade-offs)
• GraphQL implementation in production
• Familiarity with additional vector stores (Weaviate, Chroma, pgvector)
• Background in financial services or regulated industries
• Published writing or talks on AI systems architecture
• Experience with multi-modal models or agent frameworks beyond basic tool use

Pay

Competitive salary, equity, and benefits package commensurate with experience.