AI Architect

KL Software Technologies(“KLST”) iis hiring a hands-on AI Architect to lead our transformation into an agentic-AI-first product engineering organization. You will design, deploy, and operate multiple fleets of autonomous AI agents, powered by Anthropic Claude (Claude Code) and/or OpenAI Codex, that build, test, and ship software features, run product management workflows, and execute digital marketing campaigns across our flagship KLST products (learn more here www.klstinc.com/whyklstforlegal ).

This is NOT a chatbot or data-science role. You will industrialize software delivery using agentic AI developer tools – Claude Code, AI code review and optimization tools (e.g., ponytail), multi-agent product build orchestrators (e.g., gstack, which spins up CEO / PM / BA / QA / Dev agents that work together to deliver a complete product), and free / lower-cost open-source options such as NVIDIA’s agentic AI toolkits (NeMo Agent Toolkit and NIM microservices).

The Mission

Stand up and operate AT LEAST FIFTY (50) 24/7 autonomous coding and QA agents delivering product features across KLST products by the end of the year – organized so that each Senior Engineer owns and manages 5–10 agents, reviews every agent’s output, and validates results BEFORE any code is allowed to be committed.

Key Responsibilities:

• Architect, deploy, and scale a multi-agent software delivery platform using Claude Code, agentic code review/optimization tools (ponytail or similar), product build orchestrators (gstack or similar), and NVIDIA’s open-source agentic toolkits – selecting the right mix of commercial and free/open-source tooling to control cost.
• Design end-to-end agent workflows covering the full SDLC – requirements, design, coding, code review, QA automation, and release – with human-in-the-loop approval gates at every commit.
• Build coding agents that autonomously pull work items/tickets from Azure DevOps Boards, generate implementation code with Claude Code and/or OpenAI Codex to meet each ticket’s requirements, self-QA the code against positive and negative test cases, and check the validated code into the repository.
• Build product management agents that autonomously groom the backlog, draft and refine requirements and user stories, prioritize work, and generate release notes and status reporting.
• Build digital marketing agents that autonomously plan, generate, and optimize marketing content and campaigns – SEO, email, social, and web – tied to product launches and releases.
• Define and enforce the agent governance model: mandatory Senior Engineer review and validation of agent output before commits, branch protection, audit trails, rollback, and security / IP safeguards for AI-generated code.
• Enable and coach Senior Engineers to become “agent managers”, each owning and supervising 5–10 agents; build the playbooks, prompt libraries, guardrails, and evaluation metrics they use to review and validate agent output.
• Integrate the agent fleet with our Git / CI-CD pipelines (Azure DevOps / GitHub) so agents work 24/7 within guardrails across netDocShare, imDocShare, and KLapper repositories.
• Continuously measure and report agent fleet productivity, code quality, defect escape rates, and cost per feature; optimize model/toolkit selection (Claude vs. open source / NVIDIA) for cost and performance.
• Stay current with the agentic AI ecosystem (Model Context Protocol, multi-agent orchestration, agent evaluation frameworks) and continuously upgrade KLST’s Agentic Delivery Platform.

Key Qualifications:

Required Skills

• Very strong, demonstrable background setting up and operating MULTIPLE autonomous AI agents in production – designing agent architectures, orchestrating agent-to-agent collaboration, and running fleets of agents 24/7 with reliability and guardrails.
• Overall, at least TEN (10) years “hands-on” software engineering experience with a strong full-stack background (.NET / TypeScript / React or Angular / REST APIs / SQL) on Azure or AWS.
• Minimum TWO (2) years of hands-on experience building with LLMs and agentic AI developer tools – Claude Code and/or OpenAI Codex (or GitHub Copilot / Cursor / Windsurf), prompt engineering, and LLM APIs (Anthropic, OpenAI, Google, or open-weight models).
• Hands-on experience with multi-agent orchestration frameworks and toolkits – e.g., gstack, ponytail, NVIDIA NeMo Agent Toolkit, LangGraph, AutoGen, or CrewAI – including agent-to-agent workflows (PM / BA / Dev / QA agent roles).
• Hands-on