SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Arcadia is an AI-powered energy intelligence platform that helps enterprises manage utility bills, procure energy, and advance sustainability at scale. The company serves Fortune 2000 clients with unified data, AI-powered analytics, and expert advisory to drive cost savings, risk mitigation, and carbon reduction.
You will be the first engineer on Arcadia's AI Enablement team, responsible for building the company's software factory and converting manual operational processes into automated, self-reporting systems. This is a hands-on role where you'll architect and ship production-grade AI systems, establish engineering patterns from scratch, and own the full delivery lifecycle.
Key responsibilities:
- Build and ship software factory components including scaffolding, templates, and CI patterns that make the fast path the safe path
- Prototype build-versus-buy decisions and validate with working evidence
- Convert manual operational processes into instrumented automation with measurable before-and-after impact
- Instrument all systems to track accuracy, failure modes, and improvement opportunities
- Integrate AI agents and automation with Arcadia's core systems: Zendesk, Jira, Confluence, GitHub, and Snowflake
- Own the delivery path including branching strategy, code review standards, CI/CD pipelines, and release/rollback procedures
You'll work in a fully remote environment with access to co-working spaces. Bilingual fluency in English and Spanish (written and verbal) is required.
Requirements:
- Shipped LLM-backed systems into production and maintained them in operation
- Production Python experience with demonstrated discipline in typing, testing (test pyramid), dependency management, packaging, and error handling
- Owned delivery paths: branching strategy, code review standards, CI/CD pipelines, release and rollback procedures
- Expertise in building production evals: understanding what to measure, constructing datasets that expose real failure modes, evaluating judge model ROI, and setting release-blocking thresholds; ability to cite a specific eval you designed and a failure it caught
- Bilingual capability: English and Spanish (written and verbal)
Note: Visa sponsorship is not available for this position at this time, though select R&D and Data roles may offer sponsorship opportunities.
