About the Job

Build the ranking & discovery engine that turns a tsunami of AI‑generated stories into each viewer’s next obsession. Dashreels streams bite‑sized, emotionally addictive dramas—70‑minute series told through 1–2 minute episodes—all born from generative‑AI pipelines. In a world where new episodes can be spun up hourly by creators and algorithms alike, discovery is just as much an ML problem as a storytelling one. We’re hiring product‑minded engineers who can blend recommender‑system science with shipping craft to surface the perfect episode, moment by moment, for millions of viewers. You’ll own the metrics—watch‑time, completion, discovery depth—not just the systems that deliver them. This is a full‑stack role: you’ll invent Gen‑AI‑native ranking techniques, wrangle real‑time data at scale, and polish every pixel of the end‑user experience.

Responsibilities

• Design, ship, and iterate full‑stack recommendation features—front‑end widgets, back‑end services, and data pipelines—aimed squarely at boosting watch‑time, completion, and discovery depth.
• Launch to prod in week 1; improve every week.
• Own the viewer metrics by shaping ranking services end‑to‑end—from candidate retrieval and feature stores to real‑time scoring and edge delivery (Flutter + Kotlin).
• Invent Gen‑AI‑native ranking techniques—leverage multimodal embeddings, LLM‑generated summaries, and agentic diversifiers to rank and diversify freshly generated episodes without historical signals.
• Scale & optimise—build data pipelines that process billions of events, tune microservices for sub‑50 ms latency, and keep infra costs predictable.
• Instrument & iterate—build offline evaluation harnesses, run weekly A/Bs, and track hit‑rate, completion, discovery depth, and content diversity.

Requirements

We’re excited about candidates who have 3–8 years engineering experience, including at least one recommender or ranking system in production. Fluency in Python and Java/Kotlin (serving). Exposure to A/B experimentation platforms and causal ML.

Bonus: experience with LLM‑powered retrieval, agent‑style recommenders (LangGraph, AutoGen, ReAct), or integrating embeddings from multimodal generative models (diffusion, TTS, image LDMs) into ranking pipelines.

Bonus: Some mobile engineering experience

Benefits

What’s unique about Dashreels engineering:

• AI‑first catalogue – Episodes are created natively through our Gen‑AI studio, giving you metadata, scripts, and embeddings at day‑zero.
• Fast feedback loop – Short content + high engagement = rich signals for hourly retrains.
• Creator ↔ Viewer flywheel – Your rankings boost watch‑time and inform our studio which stories to green‑light next.