Applied AI Engineer

Description

Build AI features for an assistant designed to help users with conversations, daily tasks, planning, and work. The product is intended to handle long tasks, retain useful information, solve problems step by step, connect with tools, and operate reliably despite unpredictable AI responses.

This role combines machine learning, software engineering, and product development. You will improve AI models and create reliable production systems that make AI useful in real-world situations.

Main Responsibilities

• Build and deliver AI features from the model to the final user experience.
• Design and improve prompts, memory, tools, and AI agent workflows.
• Turn AI responses into clear, reliable, and consistent results.
• Find and fix problems across models, infrastructure, and user experience.
• Improve the speed, cost, and reliability of AI services.
• Create simple evaluation methods to measure AI performance in real use.
• Work closely with product managers and software engineers to solve complex problems.

Technologies

• Python
• PyTorch / JAX
• Large Language Models (LLMs), including OpenAI APIs, LLaMA, and Qwen
• Inference tools such as vLLM
• Vector databases

Requirements

• Good knowledge of machine learning and modern neural networks.
• Experience training, fine-tuning, or deploying machine learning models.
• Ability to write clean and maintainable production code.
• Comfort working across AI systems, from models to infrastructure and products.
• Strong problem-solving skills in fast-changing environments.
• Interest in delivering solutions quickly and improving them over time.

Expected Results

• AI models meet performance, speed, and reliability goals in production.
• Production problems are identified and resolved quickly with clear root-cause analysis.
• Data pipelines, model training, and inference systems are stable and easy to maintain.
• Effective collaboration with engineering, research, and product teams to deliver AI features.
• Models and systems improve based on real user feedback and measurable results.

How We Work

Great products are built by small, highly skilled teams. Team members collaborate, make decisions together, move quickly while maintaining quality, stay organized, make sound decisions, and work independently.

Interview Process

Candidates whose skills match the role may be invited to complete 3 or 4 interviews. The technical team reviews every application, and interviews are held online or in person. The process is designed to be simple and efficient.