Gen AI / Agentic Engineer

Job Summary We are looking for a GenAI / Agentic Engineer to design, build, and deploy LLM-powered applications on AWS. This role is focused on real production engineering‑APIs, RAG pipelines, agent workflows, evaluation, deployment, monitoring, and performance/cost tuning. #W2 Role Job Title: Gen AI / Agentic Engineer Location: Remote Type: W2 Contract Responsibilities Build and maintain LLM‑powered backend services using Python and FastAPI (chat, search, summarization, Q&A). Design and implement RAG pipelines end‑to‑end: ingestion, parsing, chunking, embeddings, indexing, retrieval, reranking, and grounded responses. Develop agentic workflows for multi‑step automation (tool calling, orchestration, state/memory, retries, audit logs). Deploy and support GenAI workloads on AWS using ECS/Lambda, S3, SQS, DynamoDB/RDS, OpenSearch (or vector store), and related services. Implement security and governance controls: auth, authorization, secrets, encryption, PII handling, and prompt‑injection defenses. Build evaluation and monitoring for quality, hallucination reduction, latency, and cost (test sets, regression checks, dashboards, alerts). Work across full SDLC: design docs, estimates, coding, code reviews, CI/CD, testing, release, and production support. Communicate architecture decisions clearly and explain tradeoffs (accuracy vs latency vs cost) to stakeholders. Required Skills (Point-Based) 10+ years overall IT experience with backend/API engineering and cloud deployments 2+ years hands‑on GenAI/LLM experience delivering real features (not just demos) 6+ years strong Python (core Python, clean coding, debugging, packaging) Experience with asyncio and concurrency (threads/async), plus profiling and performance tuning Comfortable with stateful/long‑running workflows: transaction handling, retries, idempotency, and failure recovery 5+ years building REST APIs / microservices, strong API design and error handling 5+ years with FastAPI (or similar) including middleware, dependency injection, background tasks Experience implementing auth/security using JWT/OAuth, RBAC, secure configuration, secrets handling Strong testing discipline using pytest (unit/integration tests, mocks, API contract testing) Proven experience building RAG systems end‑to‑end: chunking strategies, embeddings, retrieval tuning, reranking, grounding/citations Hands‑on with RAG optimization: hybrid retrieval, metadata filters, top‑k tuning, chunk tuning, reranking strategies Experience with agentic patterns: tool calling, orchestration, memory/state, structured outputs, audit trails Experience implementing guardrails: output schema enforcement (JSON), refusal handling, safety filters, prompt‑injection defenses, PII masking 5+ years AWS experience using ECS/Lambda, S3, SQS, DynamoDB/RDS (and related services) Strong AWS security fundamentals: IAM, KMS, Secrets Manager, CloudWatch logs/metrics/alarms Experience deploying LLM workloads via Amazon Bedrock (preferred) or SageMaker Strong system design: scalability, caching, rate limiting, queues, resilience/failure handling Ability to clearly explain GenAI architecture decisions and tradeoffs across accuracy/latency/cost Nice to Have LangChain / LangGraph / LlamaIndex (any) OpenSearch vector search or vector DB experience (Pinecone/Weaviate/FAISS, etc.) Docker, Terraform/CDK, CI/CD (GitHub Actions/Jenkins) Experience in regulated environments (finance/healthcare/telecom) with governance controls #J-18808-Ljbffr