Core Engineer - Software​/Applied AI; Levels

The Opportunity

We're looking for a Core Engineer focused on Software / Applied AI to build the production AI capabilities that make our edge platform scalable, reliable, and repeatable.

You'll own core pieces of our private AI platform—agentic and multi-agent systems, repeatable AI structures, evaluation and reliability mechanisms, and platform capabilities like auto fine-tuning and runtime optimization of infrastructure and models. This work operates within clear industrial production boundaries: AI can suggest and act only within well-defined limits, and we do not ship AI behavior into industrial production without evaluation, clear ownership, and a way to roll back.

You'll partner with data and infrastructure teams for requirements and feedback, then generalize learnings into platform capabilities that scale across deployments. This is a hands-on role for someone who thrives in a high-ownership setting and wants to build the infrastructure that makes real-world AI possible.

• Build and operate production AI capabilities including agentic and multi-agent workflows, tool calling, orchestration, and repeatable patterns that scale.
• Design and implement evaluation, monitoring, and quality systems that make AI behavior measurable, reliable, continuously improving, and safe in production.
• Build platform capabilities for private AI, including auto fine-tuning workflows, model/runtime optimization, and performance improvements for inference under real constraints.
• Implement safety and operational controls so AI behavior is bounded and production-ready, including policy constraints, approval workflows, auditability, and rollback mechanisms.
• Develop pragmatic interfaces and APIs that make AI capabilities easy to integrate across platform services and customer environments.
• Improve developer velocity through automation and tooling, using AI tools to accelerate implementation, tests, documentation, and iteration loops, then refining with engineering judgment.
• Partner with data and infrastructure teams to ensure the right context reaches inference and agent workflows with predictable latency, reliability, and cost.
• For senior roles: mentor engineers, review designs, and raise the technical bar across the organization.

In your first 3 months, you will have:

• Shipped at least one production AI capability (agents, evaluation, fine-tuning, or runtime optimization) that improves platform reliability, performance, or usability.
• Established a strong evaluation and rollback model for at least one AI workflow operating within industrial production boundaries.
• Earned trust through autonomy and execution—becoming a go-to owner for production AI platform capabilities.

In your first year, you will be:

• Owning major components of the private AI platform end-to-end, with clear accountability for reliability, performance, and platform adoption.
• Shipping repeatable AI structures that compress adoption cycles and scale across deployments (evaluation, guardrails, orchestration, optimization, operational playbooks).
• Driving platform evolution through product enhancements grounded in real-world constraints and measurable outcomes, with safe rollout and rollback as a default.

• 6+ years building and operating production software systems; experience shipping AI-enabled platforms or agentic systems is strongly preferred.
• Strong fundamentals in distributed systems, performance, and reliability; comfort owning production services end-to-end (e.g., Docker/Kubernetes deployments, APIs via REST/gRPC, and strong production discipline around rollout and rollback).
• Experience building evaluation frameworks, monitoring, and safety/guardrail systems that enable controlled AI behavior in production (e.g., automated eval harnesses, drift/quality monitoring, tracing, and structured telemetry).
• Strong engineering craft: clean implementations, thoughtful designs, operational clarity, and strong documentation (e.g., Python and/or Type Script/Go, FastAPI-style services, and effective testing practices).
• Comfort working in ambiguity and making sound trade-offs under real constraints (latency, cost, GPU utilization, and…