Senior/Staff Software Engineer, Agentic Platform

**Axion Ray** Senior/Staff Software Engineer, Agentic Platform; $220K–$290K base + equity; San Francisco (Hybrid)

**About Company** Axion Ray is a Series B AI platform helping manufacturers cut the cost of poor quality and improve product reliability — pulling together siloed, messy operational data into a single root-cause picture for issues like warranty claims, scrap, and rework. Customers include Caterpillar, GE, and Paccar, signed within the last few months. The company recently closed a $37M Series B (~$70M raised to date) and has no direct competitor in this category yet.

**About the Role** A senior individual contributor who architects and scales the agentic systems at the core of the Axion Ray platform, working closely with Engineering and Product leadership. The hard problem is the data — customer, service, field, telematics, warranty, and aftermarket signals that are disconnected and mostly unstructured — and the bar is high, because an agent's conclusion can drive a recall or a redesign. One of several seats on a new agentic team being built in San Francisco.

**Responsibilities** - Own the technical architecture for the agentic systems: orchestration, tool use, memory, retrieval, evaluation, observability, and the guardrails that keep agents reliable on noisy, incomplete data. - Build and harden the reasoning layer and infrastructure that take customer data from ingestion through agent-driven reasoning to root-cause insight, at the scale and reliability manufacturers require. - Establish rigorous evaluation and testing for non-deterministic systems, measuring agent accuracy against expert-validated ground truth and customer outcomes such as detection rates, investigation speed, and warranty cost. - Partner across Engineering and Product to shape the technical vision and roadmap for the agent platform. - Raise engineering standards through code review and design docs, and mentor engineers on agentic patterns and system design.

**Required Skills** A. 4+ years of software engineering experience, with recent, hands-on work building agentic systems (LLM orchestration, tool/function calling, multi-step and multi-agent reasoning, retrieval and context engineering, MCP or comparable) B. Has designed and owned complex distributed systems, ideally in data-intensive B2B SaaS C. Hands-on fluency with the modern agentic stack, plus the evaluation and observability practices needed to make non-deterministic, LLM-driven systems reliable D. Experience turning messy, heterogeneous real-world data into something a model can reason over reliably E. Technical mentorship of other engineers without formal authority

**Logistical Info** - Location: San Francisco, CA (hybrid) - Compensation: $220K–$290K base + equity

**Ideal Background** - Senior to staff-level engineer who has shipped agentic work to production and has the scars to show for it — has been burned by experiments and can tell stories about what they built, why, and what went wrong. - Opinionated about the choices they made, not just able to describe them. - Some enterprise-scale background; not purely consumer, since scale and reliability concepts differ. - Target backgrounds: teams working on Amazon Bedrock, Google Vertex, or Meta AI, which tend to bring strong operational rigor. - Strong school or company pedigree is a preference, not a requirement.

**Red Flags** - Pattern of job-hopping every six to nine months. - Experience limited to consumer-scale products with no enterprise exposure.