LLM Application Engineer

Kurz zusammengefasst

ActAI is building proactive, AI‑native applications for everyday tasks such as email, notes, and calendar management. Their first product is an agent‑first mobile communication app that aims to embed intelligence into daily conversations, errands, and workflow coordination with minimal prompting. The company's mission is to make AI a seamless part of users' lives, reducing misinformation and coordination costs while enabling high‑quality collaboration at scale. The LLM Application Engineer will create the intelligence layer that powers ActAI's AI experiences. Day‑to‑day work involves designing and shipping LLM‑driven agent workflows, building systems for reasoning, planning, memory, tool use and multi‑step execution, and turning probabilistic model outputs into reliable, observable actions. The role requires integrating large language models with APIs, databases, search services and internal tools, as well as developing prompting strategies, context engineering, structured outputs and tool‑calling techniques. You will also construct evaluation frameworks and datasets to monitor AI quality, debug issues across the full stack, from model behaviour to backend services and UX, and optimise for latency, cost and scalability. Collaboration with product and engineering teams is essential to translate ambiguous product problems into concrete AI solutions and to establish production practices for observability, tracing and continuous improvement. Candidates must have solid software‑engineering fundamentals and a proven track record of building AI‑powered applications. Hands‑on experience with large language models, generative AI or agent‑based systems is required, along with practical knowledge of designing prompts, workflows, evaluations or AI behaviour. The ability to write clean, production‑grade code and to operate comfortably across abstraction layers, from model to system to product, is essential. Strong problem‑solving skills in fast‑moving, ambiguous environments and a bias toward rapid shipping, iteration and continuous improvement are also mandatory. Nice‑to‑have expertise includes familiarity with the listed tech stack: Python programming, LLM APIs (including OpenAI‑compatible and open‑weight models), agent frameworks and orchestration systems, vector databases and retrieval mechanisms, backend services and distributed systems, as well as deep‑learning libraries such as PyTorch or JAX. Experience working on scalable, observable AI workflows and a focus on cost‑effective, predictable AI capabilities will be valued. What the role asks for: - Strong software‑engineering fundamentals - Hands‑on experience with LLMs or generative AI - Experience designing prompts, workflows or evaluations - Ability to write clean production‑quality code - Comfortable across model, system and product layers - Strong problem‑solving in ambiguous, fast‑moving settings - Bias toward shipping, iteration and continuous improvement - Nice‑to‑have: Python proficiency - Nice‑to‑have: Experience with vector databases and retrieval systems - Nice‑to‑have: Knowledge of PyTorch or JAX.