SkyLab ★ 3.1

Level: Mid to senior (5+ years building and running production software)
Location: Ho Chi Minh City, Vietnam (on- site)
Compensation: Competitive base, performance bonus, equity participation
Reports to: VP of Engineering
Language: Strong written and spoken English
The role
You will work across our new and existing products and projects. You own the features you work on from the technical design to production. You receive the product requirements, write the technical specification, plan the build, estimate and commit to your own dates, build it, and stay with it until it works in production. Nobody writes the design for you, and nobody breaks the work into tickets for you.
Our engineers use AI coding tools every day. We expect that. What we are hiring for is the judgement around the tools: deciding what to build and how before generating any code, breaking the work down, knowing when the output is wrong, and being accountable for everything that ships under your name, whoever or whatever wrote it.
We are not looking for someone who passes requirements to an AI tool and passes the result to review. We are looking for engineers who think the problem through, plan the work, and deliver it, and who use AI to do that faster.
What you will do
Design and plan

Break the build into tasks with estimates in hours and due dates, and commit to them.
Turn product requirements into a technical specification: functional requirements, acceptance criteria, data model, service boundaries, state changes and failure cases.
Raise time, feasibility and scope problems while the requirements are still being written, not after the build starts.
Present your design at architecture review, defend the decisions, and change it where the review finds real gaps.

Build and deliver

Write tests that would fail if the behaviour were wrong. Passing tests and coverage numbers are the minimum, not proof that the code works.
Keep the specification current as the design changes during the build, so QA and other engineers always work from the real design.
Review your own work before asking anyone else to, and attach the evidence when you merge: the tests, the results and what you checked.
Deploy, verify the feature works in production, and fix what breaks. Work is done when it is verified working in production, not when it is merged.
Report progress daily and raise a slip on the day you believe it, with what it affects and what you propose.
Build the feature end to end, using AI tools where they help and your own judgement everywhere.

Work with the team

Review other engineers&039; code, including AI- generated code, with the same care as your own.
Work with the product manager and QA engineer on your feature from requirements to release.
Document how the systems you build work, so the next engineer can pick them up.

Yêu cầu công việc

What we look for
In priority order:
You use AI as a tool, not a substitute for understanding. You use Claude Code, Cursor, Copilot or similar every day, and you know where they fail: invented APIs and fields, wrong assumptions about data, race conditions, missing tenant or permission checks, tests that pass without checking anything. You can explain every line you ship. You can show us how AI tools have made you faster on real work, and how you caught them being wrong.
You have strong engineering fundamentals. Data modelling, transactions and consistency, idempotency, retries and timeouts, queues and events, API design. You understand what happens when a network call fails halfway through.
You write clearly. Specifications, merge request descriptions and status updates that are specific and easy to act on. Most of our coordination is written.
You think and plan before you build. Given an unclear request, you ask the right questions, identify what is missing, and produce a plan that covers the data, the edge cases, the failure modes and how you will know it works, before you open an AI tool. You can show us specifications or designs you have written.
You have shipped and run real systems. You have built software that is running in production and been responsible for it: debugging live problems, reading logs and metrics, fixing and writing up the cause. You can walk us through systems you have built and the decisions you made.
You work independently. You take a problem from requirements to production without someone directing each step. When you are stuck, you come with options and a recommendation rather than a question alone. You deliver on the dates you commit to, and when you can&039;t, you say so early.
Requirements

Designed and owned an end- to- end data flow in production: API or webhook in, queue or sync, storage, consumer out
3+ years TypeScript/Node.js or Python as primary language, 1+ year in the other
6+ months daily use of AI coding tools on production code, with review and tests you wrote for the generated code
2+ years SQL on PostgreSQL: schema design, migrations, indexing, transactions
Strong written and spoken English
Production experience with a message queue (Kafka, RabbitMQ, SQS or similar), including idempotency and retries
Kubernetes troubleshooting of own services: kubectl, pod logs and events, rollouts
Structured logging and metrics in production (Prometheus, Grafana, Loki or equivalent)
Technical specifications or design documents you wrote that others built or reviewed against, and can show
5+ years professional software development, 3+ years on backend or full- stack services in production, with systems you designed and owned end to end
Bachelor&039;s degree in Computer Science, Software Engineering or a related field
Git with pull- request code review and CI

Nice to have

Experience building agent workflows, MCP servers or other AI tooling.
Open- source contributions, a technical blog or side projects we can look at.
Experience with workflow orchestration (for example Kestra), Redis, or vector search.
Experience with financial, accounting, e- commerce or marketplace systems, or integrations with third- party APIs.
Frontend experience with React.

What you get

Social insurance, health insurance, unemployment insurance as labor law stipulated
Professional working environment
Competitive package
Opportunity to participate in stock option program.
Premium Healthcare.
Annual leave for 15 days.
Opportunities to challenge and develop your career
Public holiday in accordance with the Vietnamese labour law

Cập nhật gần nhất lúc: 2026-10-01 14:15:05