Applied AI Engineer

Role overview Build and operate AI features that help people complete everyday communication, coordination, and productivity tasks with little or no prompting. The role covers the application layer around language models: agent workflows, context and memory, integrations, evaluation, and the production systems needed to make those features dependable.

Responsibilities - Develop and deploy product features using language models, tool calling, and multi-step agent workflows. - Create prompting, context management, memory, and orchestration systems for tasks that span multiple steps. - Build backend APIs and integrations connecting models with application services and external tools. - Design evaluations for task completion, response quality, and system reliability; use results to improve model-driven workflows. - Investigate model outputs, tool failures, and agent behavior using production logs and evaluation findings. - Improve latency, inference cost, and reliability while partnering with product and engineering teams to ship features.

Requirements - At least three years of software or machine learning engineering experience, including building AI applications. - Strong Python skills and experience developing production backend services. - Hands-on experience with language model APIs, prompting, tool use, and agent workflows. - Understanding of model capabilities and limitations, inference, and context management. - Experience evaluating AI outputs, diagnosing failures, and improving production applications. - Familiarity with APIs, databases, asynchronous processing, distributed systems, and production monitoring.

Nice to have - Experience with retrieval-augmented generation, embeddings, vector databases, or search systems. - Familiarity with open models such as Llama or Qwen and inference frameworks such as vLLM. - Model fine-tuning experience with PyTorch or JAX. - Experience building consumer AI products, particularly for messaging or productivity.