AnyBuddy.AI is a voice and text companion app for iPhone. The product direction is moving toward Vivi: a lifelong text-chat AI buddy who can plan, organize, draft, search, use phone tools the user approves, and take helpful actions through structured tools while keeping chat and action history on the phone.
This internship is for builders who want to learn how agentic AI systems are actually put together. You will not just use AI APIs. You will study and prototype parts of the framework itself: tool calls, mobile runtime constraints, prompt systems, eval harnesses, safety checks, device tools, and the product experience around an AI buddy that lives on a phone.
This is an unpaid educational internship. There is no wage, stipend, equity, or guaranteed paid role. The benefit is supervised practical experience, portfolio material, feedback, and exposure to early-stage agentic AI product development.
What You Will Learn
- How a mobile app exposes actions to an LLM through structured tool calls.
- How Vivi can use files, browser tools, calendar tools, contacts, reminders, camera, speech, and local workspace state.
- How prompt systems, skills, memory, and guardrails shape AI buddy behavior.
- How to design small evals and test harnesses for reliability, safety, and regressions.
- How a one-person company style product process works: fast iteration, careful scope, user-visible quality, and direct product feedback.
Example Project Areas
- Vivi tool reliability: improve how agent actions are described, validated, tested, and displayed.
- Mobile AI runtime: prototype workflows for agents running inside iPhone and Android app constraints.
- Camera, calendar, reminder, browser, and file tools: make phone-native actions understandable and testable.
- Guardrails and evals: create small tests for hallucination control, tool-result grounding, permission boundaries, and user safety.
- Product polish: make agentic features understandable in a consumer mobile app, especially for text-first interaction.
Who Should Apply
- You have built at least one real project, even if it was personal, school-based, open-source, or experimental.
- You can read code, debug patiently, and explain your reasoning clearly.
- You are curious about LLM agents, mobile apps, Flutter/Dart, Python, JavaScript, prompts, evals, or developer tools.
- You are comfortable working remotely and sharing short written updates.
- You want practical experience more than a polished corporate internship program.
What This Is Not
- It is not a paid internship.
- It is not a promise of employment after the internship.
- It is not a full-time unpaid developer job.
- It is not a program where interns replace paid employees or own critical production responsibilities.
How It Works
- Remote in the United States.
- Part-time, usually 5-10 hours per week.
- Typical duration is 8-12 weeks, with rolling start dates.
- Each intern gets a scoped learning project with review and feedback.
- Where appropriate, selected work may be reviewed for possible inclusion in AnyBuddy.AI or related AIbody.ai work.
How To Apply
Email support@anybuddy.ai with the subject line Agentic AI Engineer Intern application.
Include your resume or LinkedIn, GitHub or project links, and a short note answering: what part of mobile agentic AI do you most want to learn by building?
Applications are reviewed on a rolling basis through September 30, 2026.