AI Tech Lead

Inorsa•Austin, TX

About The Position

Inorsa is an AI-first automation platform transforming how mission-critical infrastructure is designed, maintained, and upgraded. We serve telecom, fiber, data centers, and more, automating everything from document ingestion to drawing generation. Our platform helps infrastructure owners build faster, safer, and more cost-effectively. We're a fast-growing, highly technical team solving complex engineering problems. Every member of our team has a direct impact on product quality, customer outcomes, and how infrastructure is delivered now and into the future. As our agentic platform matures and enterprise deployments deepen, we need a technical leader who can close the gap between AI capability and production-grade reliability. We aren't competing on general intelligence—we're winning by being the deepest AI platform in the engineering vertical. This role is the tip of that spear.

Requirements

6+ years of software engineering experience, including meaningful time building and shipping AI/ML systems in production.
Expert-level proficiency in LangGraph: persistent state management, checkpointers, complex graph topologies.
Deep experience with LangSmith or equivalent observability tooling for agentic systems.
Strong async Python and experience deploying scalable AI services in cloud environments (AWS or GCP).
Experience working with unstructured or domain-specific data: technical specs, site reports, GIS data, or similar.
Familiarity with the broader stack: TypeScript/Node.js, SQL, cloud infrastructure.

Nice To Haves

Prior experience in telecom infrastructure, structural engineering workflows, or adjacent technical verticals.
Track record building evaluation frameworks and prompt regression pipelines at scale.
Experience supporting a company through Series A or early hypergrowth phases.

Responsibilities

Architect Agentic Workflows: Design and implement multi-actor state machines using LangGraph, including loops, error-recovery paths, and human-in-the-loop checkpoints.
Own the observability and evaluation strategy using LangSmith: trace agentic paths, version prompts, and run regression tests against gold-standard engineering datasets.
Build custom evaluators (LLM-as-judge) and use traces to identify exactly where multi-step agentic chains fail in production.
Drive Vertical Domain Integration: Collaborate directly with structural and telecom domain experts to encode specialized engineering knowledge into agentic tools.
Translate 30+ years of practitioner tribal knowledge into reliable, production-grade automation.
Work cross-functionally with engineers, product managers, and field teams to ensure AI outputs meet the accuracy standards required in high-stakes engineering environments.
Set the Technical Standard: Lead by example on AI-native development: Cursor, Claude Code, GitHub Copilot, and equivalent tooling.
Automate the grunt work of coding to concentrate team energy on architecture and domain logic.
Review code, mentor engineers, and raise the bar on quality, performance, and maintainability across the AI stack.
Communicate progress, risks, and architectural tradeoffs clearly with engineering leadership and product.