Bring Agent into everyday team work
TabTin brings conversations, documents, tables, tasks, and device tools into one workspace so people and Agent can work on the same task together.
JOIN TABTIN
Join TabTinBuild a leading team harness
We are hiring across engineering, product, and design. You will work on the TabTin product and use Agent in your daily work where it helps.
01 · WHY WE EXIST
What we are buildingAgent in everyday teamwork
We use the product in our own work
The team uses TabTin and other Agent tools for planning, coding, testing, and debugging. Problems we find in daily use feed directly into product work.
02 · WHAT WE OFFER
How we workTools and collaboration
We invest in good tools, and expect everyone to review the output and own the final result.
AI TOOLS
AI tool budgetEngineers can use first-party tools such as Claude and Codex as needed, with at least US$2,000 available each day. The budget supports development, testing, and debugging; it is not a usage target.ClaudeCodexGeminiKimiDeepSeek
Agent tools in daily work
The team uses tools such as Claude and Codex for development, research, and testing. Tools help us move faster; people still own decisions, reviews, and quality.
Help build an early-stage product
TabTin is still evolving quickly. You will contribute to feature definition, technical decisions, implementation, and feedback after launch—not just receive predefined tasks.
Direct collaboration
Product, design, client, backend, and Agent Runtime work directly on the same requirement, with as little handoff overhead as possible.
Make experience reusable
When a problem repeats, we turn the solution into tests, documentation, templates, or internal tools the team can reuse.
03 · WHO WE'RE LOOKING FOR
What we valueOwnership and collaboration
The stack varies by role; these working habits matter more.
＋Who will thrive here
You already use Agent tools at work and know when they help and when to handle something yourself.
You review Agent output, catch mistakes, and take responsibility for the final result.
You care about long-term maintainability, not only whether something works today.
You can turn ambiguous problems into work that can be discussed, built, and verified.
You communicate directly and surface risks and uncertainty early.
−Who probably will not enjoy it
You only want fully specified instructions and do not want to discuss requirements or approaches.
You deliver Agent-generated work without reviewing or verifying it.
You focus only on shipping quickly, without considering maintenance or user experience.
You are unwilling to learn or use the Agent tools already used by the team.
04 · OPEN ROLES
Open roles
We are hiring in the areas below. Open a role to see its scope; level and responsibilities will be discussed based on your experience.
01BACKENDBackend engineer
Build backend systems for organizations, workspaces, permissions, resources, and Agent execution. We mainly use Java, Node.js, and Python, with a strong focus on API design, data consistency, performance, and auditing.
Apply now02FRONTEND · ELECTRONDesktop / frontend engineer
Build the Electron desktop app and web interfaces, including Agent conversations, documents, tables, browser tools, and context flowing between apps.
Apply now03MOBILEMobile engineer
Build our iOS, Android, and HarmonyOS clients, covering messaging, data sync, real-time notifications, and device integrations.
Apply now04AGENT RUNTIMEAgent Runtime / ML engineer
Work on Agent orchestration, context, memory, Skills, sub-Agents, approvals, and rollback, with a focus on reliability, observability, and safety boundaries.
Apply now05PRODUCTProduct manager
Own research, solution definition, and acceptance across Workspace, Project, Agent, Skill, permission governance, and billing. You should be comfortable using Agent for research and prototyping.
Apply now06DESIGNProduct designer
Design the product experience across desktop, web, and mobile, especially complex interactions around collaboration, Agent status, permissions, and risk prompts.
Apply now05 · HOW WE WORK
How we movea requirement forward
The details vary by role, but the basic process is similar.
- 01
Product and engineering first agree on the use case, constraints, and acceptance criteria.
- 02
We work through the data model, APIs, interactions, and risks, using Agent for research or a first draft when useful.
- 03
After implementation, the owner reviews key decisions, permission boundaries, error handling, and test coverage.
- 04
We verify locally and across relevant clients to confirm that APIs and state flows are consistent.
- 05
During review, reusable lessons are added to tests, documentation, or tools.
06 · APPLY
How to applyEmail or BOSS Zhipin
Please attach your résumé and briefly describe one piece of work you led or contributed to deeply: what you owned, how you made trade-offs, and the outcome. If you have used Agent, tell us how.
hr@larchiveai.com