Role Summary
We're looking for a Technical Architect who combines strong product engineering fundamentals with deep, hands-on expertise in agentic AI systems. This is not a research role — it's an architecture and delivery role for someone who has built and shipped real products, and who now wants to architect multi-agent, LLM-powered systems that go into production for enterprise clients. You'll own technical design end-to-end: from client conversations through architecture decisions to code reviews and production readiness.
Key Responsibilities
- Architect and oversee delivery of product-grade software systems, with agentic AI (multi-agent orchestration, tool-use, RAG, memory systems) as a core capability area
- Design and review system architecture for scalability, reliability, security, and cost — balancing agentic AI capability with production engineering discipline (observability, testing, CI/CD, versioning)
- Lead technical design for client engagements: translate ambiguous business problems into agent workflows, orchestration graphs, and system architectures
- Evaluate and select frameworks (LangGraph, LangChain, AutoGen, CrewAI, or custom orchestration) based on the problem, not fashion
- Set up evaluation, guardrails, and observability for LLM/agentic systems (e.g., Langfuse or equivalent) to ensure production reliability
- Mentor engineering teams on both classical product engineering practices and emerging agentic AI patterns
- Partner with presales/architecture leads on solutioning, estimation, and proof-of-concept builds for prospective clients
- Stay current on the LLM/agentic ecosystem (Claude, GPT, open models, MCP, agent protocols) and bring pragmatic recommendations — not hype — into client and internal conversations
- Own technical quality bar: code reviews, architecture reviews, and production readiness checks across projects
Required Skills & Experience
- Strong product engineering background: has designed, built, and shipped full-stack or backend-heavy products at scale (not just prototypes)
- Hands-on experience building agentic AI systems in production — multi-agent orchestration, tool calling, RAG pipelines, memory/state management
- Proficiency with at least one agent orchestration framework (LangGraph strongly preferred; LangChain, AutoGen, CrewAI acceptable)
- Solid grounding in software architecture fundamentals: distributed systems, API design, cloud-native deployment (AWS/Azure/GCP), containers (Docker/Kubernetes)
- Working knowledge of LLM APIs (Anthropic Claude, OpenAI, or equivalent) and prompt/context engineering at a systems level
- Experience with LLM observability/evaluation tooling (Langfuse, LangSmith, or similar)
- Strong programming skills in Python and at least one other language (TypeScript/Java/Go)
- Comfortable operating in a client-facing services environment: can explain architecture trade-offs to both engineers and business stakeholders
Preferred / Nice-to-Have
- Prior experience with AWS Bedrock or similar managed LLM infrastructure
- Exposure to MCP (Model Context Protocol) or emerging agent interoperability standards
- Experience in a technology services/consulting environment with multiple concurrent client engagements
- Anthropic or OpenAI certifications/partner-track credentials
- Prior startup or 0-to-1 product-building experience
- Has architected and shipped at least one agentic AI system into production for a client
- Has established (or improved) engineering practices around agent evaluation, observability, and reliability
- Is a trusted technical voice in client presales conversations has context menu
