Senior Software Engineer, FTVX Team (Whole World)

This role leads the Matching team responsible for entity resolution—grouping items from many providers into accurate unified clusters for Fire TV catalog discovery. You will design and evolve matching algorithms that combine ML scoring, LLM inference, deterministic rules, and quality controls to improve match decisions at scale.

• Design and evolve matching algorithms combining ML scoring, LLM-based inference, deterministic rules, external ID linking, and guard rails
• Expand production use of LLMs for match decisions including prompt strategies and evaluation
• Drive matching capability expansion to short-form, sports events, and live content
• Own end-to-end match quality (precision, recall) and customer impact
• Lead technical design, code reviews, and operational excellence

You will work on incremental matching pipelines across single-host and distributed architectures with emphasis on low end-to-end latency. The scope includes expanding production LLM usage for match inference, improving prompt strategies and evaluation, and extending matching to short-form video and live sports using temporal and event-based logic.

The ideal candidate is a senior software engineer who has led the design and evolution of entity resolution and matching systems at large scale. They have production experience using LLMs (large language models) for match inference, and can balance accuracy (precision/recall) with latency and cost while delivering high-quality clusters for end users.