Description

Candidate Skill:    Agentic AI, Python, LLM, RAG, Prompt Engineering, LangChain/AutoGen, FastAPI, Docker, Kubernetes, Vector DB, APIs, AWS/Azure/GCP, Observability
Experience:    5-7 years

Job Description:    We are looking for a Senior Agentic AI Engineer to design, build, and productionize specialized agent workflows for enterprise systems. The candidate will own end-to-end agent lifecycle management, including prompting, planning, orchestration, guardrails, tool integration, evaluation, and reliability. Initial focus will be on: Agent 1: Order intake ? discrepancy handling ? communications ? barcode Agent 2: RFS + PO creation The role requires close collaboration with business SMEs, platform teams, and other stakeholders to ensure robust, scalable, and production-ready agentic AI solutions. Key Responsibilities Agent Development & Orchestration Implement agents using Strands Agents SDK (or equivalent) with integrations across NuOrder, Freshservice, Snowflake, MIC, RFS, POWB, GXS Design prompts, workflows, and state machines for each sub-step, including planning, retries, and fallbacks Implement guardrails, hallucination detection, and low-confidence flows (HITL escalation, re-query, human confirmation) Build REST APIs and lightweight SDK wrappers (Python/TypeScript) with authentication, retries, typed schemas, and OpenTelemetry context propagation Data & Reasoning Build and operate RAG pipelines, embeddings, vector search, retrieval strategies, prompt templating, and context windows Collaborate with business SMEs to encode detection rules, templates, and PO creation logic Platform, Observability & Operations Optimize token usage, latency, throughput, and reliability Build and monitor production-grade dashboards (Grafana/Prometheus/APM) with golden signals and drill-down capabilities Integrate LangSmith, Langfuse, Phoenix, TruLens, or DeepEval for evaluation, regression suites, and dataset/version management Support feature-flag, canary, and blue-green deployments, versioned prompts/config artifacts, and SLO alerts with rollback hooks Mandatory Skills Advanced Python & Agent Frameworks – Hands-on LLM/agent development using AWS Bedrock, LangChain/LangGraph, Strands, AutoGen, or CrewAI Agent Orchestration & Reliability – Planning, retries, fallbacks, guardrails, HITL escalation, multimodal inputs, deterministic schema-driven outputs Advanced Prompt Engineering – Few-shot prompting, ReAct patterns, structured prompting, prompt decomposition, guardrails, hallucination detection RAG Expertise – Embeddings, vector search, retrieval strategies, prompt templating; experience with vector DBs like pgvector, Pinecone, or Weaviate Production Integrations – REST/GraphQL APIs, message queues (SQS/Kafka/RabbitMQ), strong error handling, idempotency Evaluation & Quality Control – LLM evaluation strategies, integration with LangSmith or Langfuse/Phoenix/TruLens/DeepEval Platform & Deployment Hygiene – Microservices (FastAPI or similar), Docker, Git, CI/CD, cloud fluency (AWS/Azure/GCP) API & SDK Implementation – REST endpoints, SDK wrappers, retries, auth, typed schemas, OpenTelemetry correlation Observability & Monitoring – Grafana/Prometheus/APM dashboards, OTEL traces, drill-down to trace level Quality & Robustness Alerts – Delta/threshold alerts, SLO breaches, on-call integrations, rollback/playbooks Business Collaboration – Work with SMEs, clear written and verbal communication Preferred Skills Frameworks: AutoGen, CrewAI, LlamaIndex, graph-based planners, toolformer patterns Vector DBs & Retrieval: Milvus, Weaviate, FAISS; hybrid search strategies Model Ops: MLflow, experiment tracking, evaluation harnesses LLM Optimization: Context caching, quantization, memory management, prompt A/B testing Caching & Cost Optimization: Embedding/result caching for production efficiency Vision & Document Parsing: PDF/image/table ingestion, OCR, schema-based JSON outputs DL Stacks: PyTorch, TensorFlow, fine-tuning (LoRA/PEFT), inference optimization Data Engineering: PySpark, Snowflake connectors, ETL production considerations Security & Compliance: PII handling, redaction, audit logs, approval flows Agent Interop: MCP client/server wiring, agent?agent handoffs HF/PyTorch integration: Local inference tests, evaluator stubs

Key Skills

Education

Any Graduate