AI Intern — Applied AI (LLMs & Vision)

Role Overview

Natoe AI is dedicated to transforming diagnostic imaging through a close collaboration between radiologists and advanced artificial intelligence. The company’s flagship product, Natoe AI Copilot, is an AI-powered radiology workflow and reporting platform that combines FDA-approved imaging models with large language models (LLMs) to generate structured preliminary reports in minutes. As an AI Intern, you’ll get hands-on exposure to real-world AI problems involving LLMs, computer vision, and multimodal AI, while working closely with the founding and technical teams.

Responsibilities

• Build and improve LLM-based generation & structuring of long-form clinical text across imaging modalities (X-ray, CT, MR, ultrasound, mammography)
• Prompt-engineer and evaluate at scale: golden sets, graders (incl. LLM-as-judge), and regression gates over thousands of de-identified reports
• Experiment with multimodal & on-premise vision-language models; benchmark competing models against each other and pick the right one per task
• Build retrieval systems (RAG) end-to-end: chunking, embeddings, indexing, ranking, and measuring retrieval quality
• Build personalization: preference memory and feedback loops that adapt output to each user

Requirements

• Strong Python and CS fundamentals (B.Tech / M.Tech, final year, or recent graduate from IIT / NIT / BITS)
• Experience shipping something with LLMs (prompts, RAG, agents, an API)
• Ability to design an eval and trust data over vibes
• Curious, ownership-driven, and high agency

Skills

• Python
• LLMs
• RAG
• Computer Vision
• Multimodal AI