Careers

Make recommendation infrastructure every business can use.

At Feiliu we use engineering discipline to turn multimodal recommendation into a standard product — explainable, controllable and compliant.

Why join

Hard problems, real impact.

  • What we do

    Turn a multimodal recommendation engine into SaaS and tooling that helps clients reach the right audience.

  • Where we are

    The core loop already runs in our mini-program, built by a core team with hands-on multi-platform experience. Multi-platform launch follows within 6–12 months.

  • What backs us

    An independent company with long-term support from strategic industry partners in technology, delivery and talent.

  • What you will own

    Recall, ranking, online features, the experimentation platform or risk control — shipped straight to production.

Open roles

Four roles we are hiring for.

Culture

How we work.

  • Ownership

    Plenty of chances to own a core module and answer for its online metrics.

  • Close to customers

    We see the effect of our work in our customers’ results, not just in offline numbers.

  • Engineering discipline

    Mature engineering and MLOps practices: observable, testable and reversible.

Growth and rewards

Grow on your terms.

  • Two career tracks

    Technical and management tracks, with transparent performance reviews.

  • Competitive pay and equity

    Competitive compensation and an equity incentive plan.

  • Diverse, measurable work

    Deep technical challenges across many scenarios, with impact you can quantify.

FAQ

Questions candidates ask us.

Is the company independent?

Yes. FEILIU TECH is operated by OBSYRA TECH PTE. LTD., an independent legal entity backed by strategic industry partners who provide ongoing support in technology, talent and delivery.

How stable is the business?

We have a clear product roadmap and customer path, focused support from our partners over the next 6–12 months, and a B2B paid model with a clear route to cash flow.

What is the product and the tech stack today?

A mini-program demo app, the algorithm SDK and API, and the console. The backend uses a mainstream stack — Go/Java, gRPC/HTTP, Kafka/Flink, vector retrieval and PyTorch — with a strong focus on observability and rollback.

What is your advantage over large platforms?

Fast delivery, customisation, explainability and compliance. We focus on putting useful recommendation capabilities in our customers’ hands, not on running a content platform that competes with them.

What does career growth look like?

Many chances to own a core module and work directly with online metrics and customer results, on technical and management tracks with transparent performance and incentives.