Sr. Agentic AI Consultant

Keys to Success:

• You have a model-first, programmability-first mindset. You think in prompts, tools, and agent graphs as naturally as in code.
• You focus on understanding data and curating powerful evaluations, including offline evals, online telemetry, and human feedback loops.
• You excel at thoroughly documenting and communicating ideas to a broad audience, including non-technical business stakeholders.
• Your attention to detail and code-craft is unparalleled. You care about what the agent actually does in the real world, not just the happy path.
• You communicate and think in a structured manner.
• You're comfortable engaging with clients, partners, peers, and anyone who has valuable input.
• You have a passion for helping others achieve success.
• You have domain expertise across multiple areas such as software development, cloud, data engineering, machine learning / LLMs, and enterprise integration.

Primary Focus:

Lead the design and implementation of agentic AI systems across client engagements, including agent orchestration, tool/function use, retrieval-augmented generation (RAG), evaluation, guardrails, and production deployment.

Key Technical Skills:

• Agent Frameworks & Orchestration: LangChain, LangGraph, CrewAI, AutoGen, Semantic Kernel, Claude Agent SDK, OpenAI Agents SDK; multi-agent coordination and hand-off patterns.
• Foundation Models & LLM Platforms: Anthropic Claude, OpenAI, Google Gemini, Meta Llama, Mistral; AWS Bedrock, Azure OpenAI, Google Vertex AI; model selection, routing, and cost/latency optimization.
• Tool Use & Integration: Function calling, structured outputs, Model Context Protocol (MCP) servers and connectors, REST/GraphQL APIs, webhooks, enterprise identity (OAuth/SAML).
• Retrieval & Knowledge: RAG architectures, vector databases (Pinecone, Weaviate, pgvector, Chroma, Milvus), embeddings, hybrid search, reranking, chunking strategies, knowledge graphs.
• Prompt Engineering: Systematic prompting patterns (ReAct, reflection, planner/executor), prompt versioning, context engineering, few-shot design.
• Evaluation & Observability: LangSmith, Langfuse, Arize, Weights & Biases, Braintrust; agent tracing, eval harnesses (golden sets, LLM-as-judge), A/B testing, human feedback loops.
• Responsible AI & Guardrails: NeMo Guardrails, Llama Guard, input/output filtering, PII detection, prompt-injection defense, policy enforcement, audit logging.
• Engineering: Python (primary), TypeScript/Node; Git/YAML, version control; test-driven development with pytest.
• LLMOps / DevOps: CI/CD, containerization (Docker/Kubernetes/OpenShift), GitOps, prompt & model versioning, documentation-as-code, secrets management (Vault/CyberArk).

Preferred Experience:

• Deploying and scaling agentic AI solutions in production enterprise environments.
• Designing multi-agent systems with planner/executor, critic, and human-in-the-loop patterns.
• Architecting model-agnostic orchestration layers and abstraction frameworks across providers.
• Building agent evaluation and observability pipelines, including eval-driven development workflows.
• Integrating agentic workflows with enterprise systems (ServiceNow, Jira, Salesforce, M365, ITSM platforms).
• Fine-tuning, LoRA/adapters, distillation, and model-routing strategies for cost and performance.
• Enterprise AI governance, responsible AI frameworks, and compliance (NIST AI RMF, EU AI Act awareness).

Soft Skills:

• Excellent documentation and knowledge-sharing skills.
• Strong cross-team collaboration and communication.
• Self-driven problem solver; comfortable with ambiguity, which is a must in a rapidly evolving AI landscape.
• Ability to convert business requirements into agentic AI solutions with clear ROI.

Qualifications:

• Ability to perform concurrent tasks in complex environments under adjusting priorities.
• Ability to communicate and modify approach, language, and style to different audiences, including C-suite executives.