AI Engineer, Internship – Intelligent Question Bank Platform

Das ist der Job

Responsibilities You will architect and implement the core AI pipeline that powers Accel's test creation system.

Darum lohnt es sich

Work closely with the founder to design and build an AI-powered content generation system from the ground up. You'll contribute to meaningful parts of the product end‑to‑end from how the system ingests and understands source material, to how it produces and validates outputs, to how instructors interact with and review what the system generates.

On the engineering side, you'll build and iterate on LLM-driven pipelines, work with retrieval and embedding techniques to ground outputs in real source material and develop backend services and APIs that tie everything together.

Beyond pure coding, you'll be expected to think about output quality and building evaluation steps, catching failure modes, and improving the system based on real instructor feedback. You'll research new tools and techniques as the AI space evolves and bring relevant ideas directly into the product.

This is a generalist role at an early‑stage product where you'll wear multiple hats, work with ambiguity, and have direct input into how things are built.

Requirements Strong foundation in software engineering: data structures, APIs, system design Proficiency in Python (primary language for AI/ML pipeline work) Experience with REST APIs and at least one database (PostgreSQL preferred) Ability to work independently, ask sharp questions, and iterate fast Strong debugging and problem-solving instincts Demonstrated side projects or shipped code (GitHub portfolio required) Genuine interest in AI systems and education technology Direct experience with LLM APIs: OpenAI, Anthropic Claude, or Google Gemini Hands‑on experience with RAG systems: embedding models, vector databases (Pinecone, Weaviate, pgvector, Chroma) Familiarity with prompt engineering techniques: few‑shot prompting, chain‑of‑thought, structured JSON outputs Experience with NLP pipelines: text chunking, tokenization, semantic search Knowledge of LaTeX syntax and math rendering libraries (MathJax, KaTeX) Experience with image generation APIs or SVG programmatic generation Familiarity with AI evaluation frameworks or automated test harnesses for LLM outputs Cloud platform experience: AWS, GCP, or Vercel for deployment Experience with job queues: Celery, Bull, or similar Exposure to educational content standards or psychometrics is a bonus #J-18808-Ljbffr