IT Intern — AI Service Desk Assistant

Lyten•San Jose, CA

2d•$25 - $32

About The Position

The IT Intern will partner with the IT Service Desk team to design, build, and pilot an AI-powered IT Service Desk Assistant. This assistant will use large language models (LLMs) and retrieval-augmented generation (RAG) against Freshworks ticketing data and the IT knowledge base to auto-classify tickets and recommend resolutions to L1 analysts. This is a hands-on, project-based internship with a defined deliverable, an executive sponsor (Senior Director, IT & Security), and a dedicated technical mentor. The intern will own the prototype end-to-end, including data exploration, model and retrieval design, pilot execution, measurement, and a production-readiness recommendation.

Requirements

Currently enrolled in a Bachelor's or master’s program in Computer Science, Data Science, Information Systems, AI/ML, or a related field; rising junior, senior, or graduate student preferred.
Proficiency in Python or like language.
Working knowledge of LLM APIs (OpenAI, Anthropic, or equivalent) and basic prompt engineering.
Familiarity with RAG concepts.
Git and version control fundamentals.
Strong written and verbal communication; able to explain technical work to non-technical stakeholders.
Self-directed, organized, and able to scope and ship a deliverable on a fixed timeline.

Nice To Haves

Prior coursework or project work in NLP, information retrieval, or applied machine learning.
Experience with ITSM platforms (Freshworks, ServiceNow, Jira Service Management) or general familiarity with IT service desk operations.
Familiarity with evaluation frameworks for LLM applications (RAGAS, LangSmith, or equivalent).
Basic understanding of data privacy and security best practices when working with internal company data.

Responsibilities

Extract, clean, and analyze ~12 months of Freshworks ticket data to baseline volume, categories, MTTR, and resolution patterns.
Build a ticket classification and routing model using a commercial LLM API.
Design and implement an RAG pipeline against the Lyten IT knowledge base using a vector store.
Iterate on prompt design, retrieval quality, and evaluation metrics.
Run a structured pilot with a small group of L1 service desk analysts and measure impact against baseline KPIs.
Document the architecture, runbook, and a recommended production path.
Present findings and recommendations to IT leadership at the end of the engagement.