About the job
Paraform is hiring an Applied AI Engineer in San Francisco. This role focuses on building and deploying AI systems that directly serve real users. The ideal candidate brings 2-5 years of experience, a strong grasp of modern LLM-based technologies, and a track record of turning advanced models into reliable product features. Success in this role depends on sound product sense and the ability to weigh trade-offs between LLM and traditional machine learning approaches. Experience with LLM-powered applications, retrieval systems, agentic workflows, or automation is valuable. Familiarity with classic ML techniques, such as ranking, recommendation, or classification, will help in designing hybrid systems that balance performance, cost, and reliability.
What You Will Do
- Design and build AI systems to improve matchmaking, ranking, and automation in the Paraform marketplace.
- Develop LLM-driven features, including retrieval pipelines and agentic workflows, to streamline recruiter and company interactions.
- Own systems end-to-end: from data pipelines and model design to deployment, monitoring, and iteration in production.
- Work closely with product managers, ML engineers, and full-stack teams to deliver AI capabilities that shape marketplace outcomes.
- Create evaluation frameworks to measure real-world performance, reliability, and business impact, not just offline metrics.
- Set best practices for building and maintaining production AI systems, balancing model quality, cost, latency, and maintainability.
- Advance the integration of AI into product experiences across the platform.
What We Look For
- 2-5 years of experience at an AI-focused startup (Series A through D).
- Background working on products with a broad user base, beyond single-enterprise deployments.
- Proficient in Python and Typescript.
- Experience developing agentic systems that drive measurable business or user outcomes.
- Comfort with ambiguity and building in 0 to 1 environments.
- Ability to communicate technical trade-offs clearly to non-technical stakeholders.

