Applied AI Engineer

About the Role

As an Applied AI Engineer, you will turn raw model capabilities into reliable, real-world product behavior. You will own problems end-to-end—from shaping model behavior and building the systems around it to ensuring consistent performance in production. Sitting at the intersection of machine learning, systems engineering, and product development, this role focuses on making AI genuinely useful and reliable for users in real-world scenarios, not just in demos.

Core Focus & Responsibilities

• End-to-End Delivery: Build and ship AI-powered features end-to-end, spanning model systems, backend infrastructure, and user-facing experiences.
• Workflow Design: Design, test, and iterate on prompts, persistent memory, external tools, and agent workflows.
• Output Reliability: Transform raw model outputs into structured, reliable, and predictable product behaviors.
• Full-Stack Debugging: Diagnose and resolve issues across the entire stack, including models, orchestration layers, infrastructure, APIs, and UX.
• System Optimization: Optimize AI-powered features for latency, cost efficiency, scalability, and production reliability.
• Evaluation Frameworks: Develop practical evaluation frameworks and metrics to measure AI performance in real-world scenarios.
• Cross-Functional Collaboration: Partner closely with product, engineering, and research teams to translate ambiguous problems into reliable, production-ready systems.

What We’re Looking For

• ML Foundations: Strong theoretical and practical foundation in machine learning and modern neural network architectures.
• Hands-On Experience: Practical experience training, fine-tuning, evaluating, or deploying machine learning models.
• Production Code: Ability to write clean, maintainable, and production-quality code.
• Cross-Layer Comfort: Comfortable working across multiple abstraction layers, from model → infrastructure → product.
• Problem-Solving & Mindset: Strong problem-solving skills and the ability to operate effectively in ambiguous, fast-moving environments, with a strong bias toward shipping, experimentation, iteration, and continuous improvement.

Tech Stack

• Languages & Frameworks: Python, PyTorch / JAX
• LLMs: OpenAI-style APIs, LLaMA, Qwen, and other open-source or proprietary models
• Inference & Serving: vLLM or similar model-serving frameworks
• Data Stores: Vector databases

Expected Outcomes

• Target Performance: Ensure production ML systems meet expected accuracy, latency, reliability, and quality targets.
• Effective Resolution: Identify, diagnose, and resolve production issues quickly while addressing underlying root causes.
• Maintainable Systems: Build robust, reproducible, scalable, and maintainable data pipelines, training workflows, and inference systems.
• Signal-Driven Iteration: Continuously improve models and systems based on real-world usage signals, evaluation results, and measurable performance gains.
• Seamless Collaboration: Work effectively with engineering, product, and research teams to deliver reliable, high-impact ML-powered features.

How We Work & Application Process

We are a small, world-class team with a high talent density. We operate at a rapid pace while balancing high-quality engineering craftsmanship with continuous learning, experimentation, and iteration.

• Interview Process: If there appears to be a mutual fit, we will schedule 3–4 interviews with members of the technical team, conducted virtually and/or on-site. We prioritize transparency and efficiency and aim to make timely, well-informed decisions throughout the process.