Nearby AI Internship Program - Engineering Track

NewsBreak•Mountain View, CA

20h

About The Position

NewsBreak, a leading Content Intelligence platform, is seeking interns for its Nearby AI division. Nearby AI is building a trust-first marketplace for local home services, aiming to provide homeowners with clarity on problems, costs, and provider suitability before connecting them. The role involves contributing to the end-to-end personalized provider-recommendation pipeline, including user understanding with LLMs, provider retrieval and ranking, backstage agent workflows, multi-source knowledge integration, and evaluation. This is a hands-on role working with engineers and product managers from design through launch.

Requirements

Pursuing a bachelor's degree or higher in Computer Science, AI, or a related field.
Proficient in Python or Java with solid data structures and algorithms.
Hands-on experience developing LLM applications or agents (RAG, agent workflows, evaluation harnesses).
Understanding of search, recommendation, or ML fundamentals.
Strong problem-solving and communication; initiative to own results.

Nice To Haves

Embeddings, hybrid search, reranking, learning to rank, or personalized matching.
Elasticsearch, MongoDB, PostgreSQL, vector databases, or data pipelines.
Shipped products, open-source contributions, or independent projects showing individual impact. Research and personal projects count; GitHub links and demos welcome.

Responsibilities

Help build the personalized provider-recommendation pipeline end to end: user understanding with LLMs, provider retrieval and ranking, backstage agent workflows, multi-source knowledge integration, and evaluation, working with engineers and product managers from design through launch.
User understanding and personalization: use LLMs to extract service requirements, preferences, and constraints from conversations, profiles, and context.
Provider retrieval and ranking: combine semantic search, recommendation algorithms, and provider features (capabilities, coverage, transaction-backed reputation, user fit) to improve relevance and coverage.
Agent workflows: build backstage pipelines connecting intent understanding, retrieval, provider comparison, and recommendation decisions, with explanations a user can check.
Multi-source data and knowledge integration: entity matching across business information, reviews, credentials, and service coverage.
Evaluation and delivery: build evaluation datasets, analyze failure cases and user feedback, improve quality, reliability, and latency; ship and iterate.