全职 Founding Engineer, Platform - AI Systems 工作, 薪水 up to SGD 10,000, GENGIS AI PTE. LTD. Central Region (Singapore) 公司招聘中 - Ricebowl

The role

Gengis AI is building applied AI for the operational layer of film, television, and media production, the hard work that happens after generation: turning creative material into structured workflows that production teams can review, refine, approve, and use. We are not building a foundation model. We are building a model-agnostic platform that makes AI useful inside real production environments. You will own the core systems, AI orchestration, evaluation workflows, and product infrastructure that make this possible. You will work directly with the founders and early production users to ship the first durable version of the platform. In the first phase, your focus will be a reliable workflow that helps production teams move from creative input to practical planning and visual development outputs. This is not a narrow backend role and not a pure ML research role; you should be able to reason across systems, product, applied AI, infrastructure, data, and user-facing execution.

What you'll do

• Core platform architecture and backend systems
• AI orchestration across text, image, and future video workflows
• Reliable async workflows for long-running AI and production tasks
• Evaluation systems that make output quality measurable, repeatable, and improvable
• Human-in-the-loop review flows where users can refine, approve, reject, and iterate
• Cost, reliability, and performance controls for AI-powered workflows
• Data handling, permissions, auditability, and security fundamentals
• Cloud deployment, CI/CD, observability, and operational discipline
• Engineering standards and early technical hiring as the team grows

What we're looking for

• 7+ years building and shipping production software, with real ownership of systems taken from zero to live use
• Strong backend and systems architecture ability: APIs, databases, stateful services, async jobs, queues, and cloud infrastructure
• Experience building data-heavy, workflow-driven, or human-in-the-loop products
• Experience building with LLMs, generative AI APIs, diffusion/video models, or AI orchestration systems in production or near-production environments
• Practical judgment around model failure modes, evaluation, cost, latency, reliability, and user trust
• Comfortable designing evaluation workflows and quality metrics, not just shipping demo prompts
• Full-stack enough to ship end-to-end early, when the team is small
• Cloud and infra fluency: deployment, CI/CD, inference basics, observability, and security fundamentals
• Clear communication with product, creative, and non-technical users
• Ability to set technical direction, make tradeoffs, hire well, and lead without over-engineering

Nice to have

• Experience with AI evaluation, model routing, provider abstraction, or cost/performance benchmarking
• Experience with MLOps or data pipelines: versioning, reproducible evaluation, and dataset governance
• Experience with media, film, creative tooling, workflow platforms, or collaborative review tools
• Experience with Postgres, cloud GPU workflows, or event-driven infrastructure
• Early-stage startup experience where you built the first durable version of a product

You'll thrive here if

• You want to own the technical layer that turns a strong vision into a working platform
• You build for professionals who need AI to be useful, controllable, reliable, and commercially practical, not just impressive in a demo
• You can ship the core system largely hands-on while setting architecture for the team that follows
• You make pragmatic tradeoffs between speed, quality, cost, and runway
• You want senior ownership from day one, with a path toward Head of Engineering as the company scales

How to apply

Send your application to ************* and include:

• CV and links to shipped work or code
• A short note on one complex system you built or owned: what was hard, what you measured, what tradeoffs you made, and what you would do differently now