Focused agent architecture
Design capabilities that remain understandable as instructions, data, documents, Memory, Skills, and tools evolve independently.
Careers at AutronAI
AutronAI is building a platform where focused Agents can work with data, documents, Memory, Skills, MCP tools, scheduled Automation, and supported communication channels. We care about the unglamorous details that turn a promising model into a dependable product: clear boundaries, inspectable behavior, careful permissions, and an interface people can operate.
We would rather be precise about that than publish placeholder roles. This page explains the problems we work on, how we approach them, and what future teammates can expect. If the direction fits your experience, you can introduce yourself for future consideration.
Language models are only one part of an operating AI system. The surrounding product must decide which information is relevant, which tools are available, how capabilities are composed, where interactions happen, when scheduled work runs, and how an operator understands the outcome.
That is the product surface AutronAI is building. It spans frontend product engineering, backend services, AI orchestration, data access, integrations, automation lifecycle management, observability, and the writing required to make complex behavior understandable.
Design capabilities that remain understandable as instructions, data, documents, Memory, Skills, and tools evolve independently.
Connect MCP tools and provider integrations without treating every available action as an implicit permission.
Make scheduled execution, run state, errors, cost, latency, and delivery outcomes visible enough to operate.
Turn technical contracts into clear workflows for people who should not need to understand internal orchestration.
We describe what the system actually supports and verify behavior against code, tests, and runtime evidence.
We prefer focused increments that can be tested and understood over broad rewrites that hide risk.
Credentials, permissions, user input, and cross-service boundaries deserve explicit treatment.
Names, interfaces, errors, documentation, and visual hierarchy should help someone form the correct mental model.
AutronAI is a full-stack system. Future roles may sit in one discipline, but strong work requires curiosity about the adjacent contracts.
You can move between an ambiguous product problem and a concrete implementation without inventing unsupported requirements. You communicate tradeoffs early, treat tests as evidence rather than ceremony, and leave the codebase easier to reason about.
You do not need to agree with every existing decision. You do need to inspect the system before replacing it, explain why a change is worth its cost, and care about the experience of the person operating what you build.
No public roles are listed on this page at the moment. We will publish concrete scopes here when an opening is ready.
Yes. Send a concise note with the kind of problems you work on, relevant examples, and links that help us understand your contribution.
The contact route below is intended for people interested in contributing directly. Unsolicited bulk candidate submissions are not requested.
The public company contact is in Ho Chi Minh City. Location and working arrangement will be stated on each future role rather than implied here.
Introduce yourself with context: the systems you have worked on, the decisions you owned, and the AutronAI problem area that interests you.