AI Software Engineer

About the role

We are hiring an AI Software Engineer to own two things: the data management platform, and the bot workflows that automate business processes across the Intelligent Platform (UDP).

You will keep the site-selection map, data APIs and crawling pipelines running and improving. You will also build agents and bots that take a request in a chat tool and carry it through to a result: a report, a data load, a ticket, or a reviewed pull request.

What you'll build

• You will design and ship bot workflows that turn a business request into a finished, checked result with as little manual work as possible.
• Build agents that read data, call internal APIs and reply in chat tools, with files and screenshots when useful.
• Extend our agentic engineering platform, which files work items, dispatches coding agents, opens pull requests and gates every merge on review.
• Automate recurring business processes end to end: data requests, report delivery, partner data loads, alerting, ticket triage and follow-up.
• Measure each workflow: cost per run, success rate, time saved, and how often a human had to step in.
• Keep every bot secure: SSO, least-privilege access, workspace membership checks, secret handling and audit logs.

Must have

• 3+ years building and running production software, with both backend and frontend work shipped.
• Strong engineering fundamentals: data structures, API and database design, SQL, testing and debugging. We care more about how you reason than which language or framework you know.
• Able to work across the stack, from backend services to a modern web frontend, and to review data-heavy UI code whether you or an agent wrote it.
• JD — AI Automation Software Engineer
• Product sense: you can turn a vague business need into clear requirements and acceptance criteria, ask the right questions of business users, and judge whether what was built actually solves their problem.
• Daily use of agentic coding tools, with the judgement to review, test and correct their output.
• Hands-on experience building LLM applications or agents: tool use, prompts, evaluation, and controlling cost and failure modes.
• Cloud experience: deploying and running services, API gateways, identity and SSO authentication.
• CI/CD and containers, and comfort owning deploys and production incidents.
• Able to take over an unfamiliar codebase from its git history and documentation, and write clearly for the next person.

Nice to have

• Modern data platforms: scheduled jobs, data catalogs and governance, SQL warehouses.
• Geospatial work: spatial indexing, map layers and POI data. Low-code automation and BI: workflow automation, embedded reports and report delivery. Infrastructure as code. Chat bots and agent tool-integration protocols.
• Retail or FMCG domain knowledge: store networks, site selection, SKU analytics. Vietnamese and English for working with business users and documentation.

How we work

Every change goes through a pull request with automated checks and review. Coding agents already open many of our pull requests, so you will spend part of your time writing clear acceptance criteria and reviewing agent output, not only writing code.

First 90 days

1. By day 30: you run the hand-over. You can deploy each inherited repository, you own its on-call, and you have written a short runbook for every pipeline and API.

2. By day 60: you have shipped your first production bot workflow end to end, with cost and success metrics, and fixed the top recurring errors in the inherited systems.

3. By day 90: you own a backlog of processes to automate, ranked by time saved, and at least one business team uses a workflow you built every week.