Junior AI Search Engineer

Shade.inc•New York, NY

6d•Onsite

About The Position

Shade is scaling rapidly, having built out the combined tech of Frame.io (acquired by Adobe for $1.275B) and LucidLink, augmented with proprietary AI search and labeling. We provide critical infrastructure for post-production houses, creative agencies, sports teams, and internal media teams at large companies. Our customers include Salesforce, Snowflake, A24, the Boston Celtics, HelloFresh, Deloitte, and Motorola. We are currently ingesting 20-30% of Dropbox's daily data, amounting to 50TB per day, 10 million minutes of video per month, and hundreds of millions of assets. The company is experiencing 200% QoQ growth with 120% NRR and has recently closed a $14M Series A round backed by prominent investors. Our core focus areas include solving search at petabyte scale with a multi-modal engine that understands user intent beyond keywords, addressing the unreliability of large-scale data transfer (hot to archive storage, cloud to cloud, camera to editor), reimagining version control for creative media teams, and creating a unified layer for storage that integrates with various tools like project management, AI generators, and editing software.

Requirements

0-2 years of full-time engineering experience, or strong internship, research, or project work that shows you can ship production code.
Solid Python fundamentals — comfort reading and writing backend code, not just notebooks.
Some exposure to LLM-powered systems: RAG, embeddings, semantic search — whether through work, coursework, or a side project you’re proud of.
Curiosity about how retrieval systems work and why they fail.

Nice To Haves

Pre-Series B startup experience is a plus, but not required.

Responsibilities

Grow into owning the search layer on top of the existing foundation.
Work directly alongside the engineering team to learn how production AI search systems are built and operated at scale.
Take on ownership of the search layer as you ramp up.
Work on the indexing and retrieval pipeline, including improving LLM re-ranking, evaluating retrieval quality, and iterating on chunking and embedding strategies.
Make architectural decisions with consideration for cost consequences.
Perform real work on a live product from day one, as customers depend on it.
Work on integrating external data sources and information outside of the platform to ensure the system plays well with a customer's entire business.