Technical Engineer

Build the pipeline that turns a stated goal into a finished film. The system orchestrates agents against long-running generation jobs across several model providers, which means most of the interesting work is in reliability, cost control and the seams between services rather than in any single clever function.

• Build and maintain backend services in Python — Flask, Celery, Postgres
• Integrate video and language model providers whose APIs are asynchronous, rate-limited and inconsistent with each other
• Own the generation pipeline end to end: submission, polling, retry, storage, delivery
• Make cost visible and bounded — every generated second is billable and someone is paying for it
• Instrument the system so a failed film can be explained without reading logs by hand
• Work directly with Video Engineering on output quality problems that are not code problems
• 4+ years building production backend systems in Python
• Real experience with asynchronous job pipelines — queues, workers, idempotency, retries
• Comfort with Postgres beyond ORM defaults: query plans, indexes, connection pooling
• Experience integrating third-party APIs that fail in interesting ways
• A bias toward diagnosability — you have been on the wrong end of a system that could not explain itself
• Able to overlap at least four hours daily with Singapore time