Lead - AI Architect

Lead - AI Architect (5+ Years)
Location: On Site | Gurugram, India
Shift Timing: 01:00 PM to 10:00 PM IST (Monday – Friday)
Employment Type: Core Team Contract Role
Company Weblinks: https://allneurons.ai/ https://pragyaa.io/ https://legalgraph.ai/
JOB DESCRIPTION
A fast-moving AI product company is hiring an AI Architect to own the architecture behind our Claude-powered products. This role is for someone with deep, hands-on expertise in Claude, including Anthropic's model family, its agentic tooling (Claude Agent SDK, Model Context Protocol, tool use, extended thinking), and the practices for building reliable, auditable production systems on top of it. Official Claude certification is required, since this person will set the technical standard the rest of the engineering team builds against.
The engineering environment is centered around solving high-value workflow problems through AI, including contract review, compliance automation, document intelligence, and back-office operations. We would describe our platform as one that automates contract review, compliance workflows, and financial analysis, with an emphasis on transparency, auditability, and enterprise use cases, which makes architectural rigor, not just prototyping speed, essential to this role.
Responsibilities
Architect end-to-end AI systems on Claude, including agentic workflows, RAG pipelines, tool use, and multi-agent orchestration, for contract review, compliance automation, document intelligence, and financial analysis.
Define the technical standards for how the org builds with Claude: model selection across the Claude family, context and prompt architecture, evaluation frameworks, guardrails, and cost/latency tradeoffs.
Own the reliability, scalability, auditability, and transparency of Claude-based systems running in production for enterprise customers.
Evaluate and integrate Claude's ecosystem, including the Claude Agent SDK, Model Context Protocol (MCP), tool use/function calling, extended thinking, and prompt caching, into the product architecture.
Partner closely with founders, product, and engineering to translate ambiguous product goals into concrete, well-architected AI solutions.
Set and enforce best practices for AI observability, safety, and governance across the platform.
Act as the organization's go-to Claude expert: mentoring engineers, reviewing designs, and staying current with Anthropic's model and product releases.
Stay hands-on: prototype and validate architectural decisions directly in code when needed, not only on paper.
Requirements
Holds an official Anthropic Claude Certification: Claude Certified Architect (Professional or Foundations) required; Claude Certified Developer – Foundations is a strong plus.
5+ years of experience in software or AI/ML architecture, with recent, hands-on depth in LLM-based and agentic systems.
Demonstrated production experience architecting RAG pipelines, agentic workflows, and multi-step tool-use systems specifically with Claude.
Strong backend engineering foundation (Node.js and/or Python) sufficient to prototype and stress-test architecture decisions hands-on.
Experience designing for auditability, transparency, and regulatory compliance, directly relevant to our contract-review and compliance domain.
Excellent communication skills, able to explain architecture tradeoffs to founders, product managers, and non-technical stakeholders.
High ownership and comfort operating in a fast-paced startup environment.
Ability to work regularly in the PDT-aligned window (typically 1 PM – 10 PM IST, Monday–Friday) to collaborate with US-based stakeholders and teams.
Nice to Have
Additional Anthropic/Claude Partner Network training or certifications beyond the required credential.
Experience building or deploying custom MCP servers.
Familiarity with LLM evaluation tooling (e.g., promptfoo, LangSmith) and vector databases.
Prior experience in legal tech, compliance, or fintech product environments.

Work arrangement
No