Agentic AI Engineering Intern

Deion is a company providing a unique alternative for YouTube thumbnail design, powered by a proprietary algorithm. They are seeking an Agentic AI Engineering Intern to design and implement autonomous systems, leveraging AI technologies to solve complex tasks and integrate diverse data sources.

Responsibilities:

• Design and implement custom autonomous agent architectures using Python and Google Cloud infrastructure
• Leverage Gemini CLI and Claude Code to accelerate development cycles and orchestrate complex agentic task execution
• Build bespoke reasoning loops and decision-making logic to solve open-ended problems without relying on pre-built low-code platforms
• Integrate diverse data sources and external services into a cohesive system via robust, Python-based middleware
• Debug and refine non-deterministic agent behaviors to ensure reliability, safety, and alignment with user goals
• Architect modular systems where agents can autonomously select and use the correct tools for a given task
• Collaborate with cross-functional teams to define, implement, deploy, and scale AI solutions within a cloud-native environment
• Document system designs, source code, workflow designs, research, development plans, and engineering standards to enable solution modularity, reproducibility, portability, collaboration, and knowledge sharing
• Collaborate with teammates using Slack, Notion, Google Meet, and/or other software

Requirements:

• Completion of a demo project
• At least 10 hours of synchronous meeting availability between 1:00 PM and 5:00 Central Time on weekdays, for the next 6 months
• Comfortable working 100% remotely as part of a distributed team
• Excel at designing autonomous systems that can reason through complex tasks using LLMs
• Work products demonstrate a high level of attention to detail and the ability to see the big picture
• Aptitude, mindset, and ability to learn quickly
• Design and implement custom autonomous agent architectures using Python and Google Cloud infrastructure
• Leverage Gemini CLI and Claude Code to accelerate development cycles and orchestrate complex agentic task execution
• Build bespoke reasoning loops and decision-making logic to solve open-ended problems without relying on pre-built low-code platforms
• Integrate diverse data sources and external services into a cohesive system via robust, Python-based middleware
• Debug and refine non-deterministic agent behaviors to ensure reliability, safety, and alignment with user goals
• Architect modular systems where agents can autonomously select and use the correct tools for a given task
• Collaborate with cross-functional teams to define, implement, deploy, and scale AI solutions within a cloud-native environment
• Document system designs, source code, workflow designs, research, development plans, and engineering standards
• Collaborate with teammates using Slack, Notion, Google Meet, and/or other software
• Strong proficiency in Python with a focus on writing clean, scalable, and highly modular code
• Practical experience deploying and managing applications within the Google Cloud ecosystem
• Familiarity using advanced developer tools like Gemini CLI or Claude Code to solve technical challenges
• Demonstrated systems thinking skills with the ability to map out intricate logic flows and manage state across complex processes
• Ability to decompose high-level, ambiguous objectives into a series of small, autonomous, and verifiable tasks
• Experience using Git-based version control to manage codebases within a collaborative team setting
• Excellent communication skills for explaining technical trade-offs and architectural decisions to both technical and non-technical peers
• A high degree of adaptability and the 'engineer's intuition' required to troubleshoot probabilistic systems
• Effective modular thinking skills evidenced by the ability to break down large, intricate undertakings into a series of smaller self-contained tasks
• Excellent written and verbal communication skills
• Emotional intelligence and coachability