Engineering Internship - AI/Data

iudexnc; Zürich; Vor Ort; CHF 28.000 - 50.000; Vollzeit

Benefits dieser Stelle

Direct impact

Learning & mentorship

Breadth of work

Founder access

ESOP potential

Zusammenfassung

INC (Iudex Non Calculat) is a pre-seed AI startup based in Zurich, building infrastructure that automates coverage checks and case handling for legal insurance companies. Legal insurance is a €16.8B European market where every case still requires 15-45 minutes of manual review.

We compress that to under 2 minutes. You're in the final stretch of your studies and seek a hands-on internship to build real AI systems that ship to production.

Qualifikationen

• End of studies: Master’s or near graduation.
• Degree in CS/DS/Math or equivalent experience.
• Solid programming skills in TypeScript/Node.js.
• Comfort with data structures, algorithms, and messy real-world data.
• Working knowledge of SQL (PostgreSQL preferred).
• Interest in applied AI/ML (LLMs, embeddings, retrieval, agentic workflows).
• Strong ownership and clear English communication.
• Availability for full-time internship, ideally 6 months; on-site in Zurich at least twice per week.

Aufgaben

• Data pipelines: Build and improve ingestion and processing for policy documents, legal texts, and case descriptions.
• AI-ready structuring: Chunking, metadata extraction, clause mapping, and schema design for reliable downstream use.
• LLM agent workflows: Prototype and implement agentic LLM workflows for coverage analysis with accuracy and cost awareness.
• Search & retrieval: Assist with indexing, embeddings, hybrid retrieval, and relevance tuning for legal content.
• Evaluation & quality: Create eval sets, run experiments, analyze failures, and ship improvements.
• Engineering fundamentals: Write clean TypeScript, add tests, and improve observability for pipelines and services.
• Collaboration: Work closely with the founder and domain experts; document decisions and share learnings.

Kenntnisse

TypeScript/Node.js

SQL

English

Data structures

Algorithms

Communication

Ausbildung

CS/DS/Math degree (in progress)

Tools

Docker

CI/CD basics

Azure

Elasticsearch

Vector databases

PDF tooling / OCR

Jobbeschreibung

You're in the final stretch of your studies (in particular graduated Master's student or close to graduation) and you want a hands‑on internship where you build real AI systems that ship to production. You enjoy working with messy data, writing solid code, and running experiments to improve LLM‑powered workflows. You learn fast, take ownership, and like collaborating closely with a small team.

INC (Iudex Non Calculat) is a pre‑seed AI startup based in Zurich, building infrastructure that automates coverage checks and case handling for legal insurance companies. Legal insurance is a €16.8B European market where every case still requires 15‑45 minutes of manual review. We compress that to under 2 minutes. We're backed by a paid pilot with a Swiss insurer and a founder with 120+ B2B customers at his previous AI legal tech company. We work at the intersection of LLM agent systems, document understanding, and domain‑specific reasoning — where accuracy, traceability, and auditability matter.

About You

You're in the final stretch of your studies (in particular graduated Master's student or close to graduation) and you want a hands‑on internship where you build real AI systems that ship to production. You enjoy working with messy data, writing solid code, and running experiments to improve LLM‑powered workflows. You learn fast, take ownership, and like collaborating closely with a small team.

About INC

INC (Iudex Non Calculat) is a pre‑seed AI startup based in Zurich, building infrastructure that automates coverage checks and case handling for legal insurance companies. Legal insurance is a €16.8B European market where every case still requires 15‑45 minutes of manual review. We compress that to under 2 minutes. We're backed by a paid pilot with a Swiss insurer and a founder with 120+ B2B customers at his previous AI legal tech company. We work at the intersection of LLM agent systems, document understanding, and domain‑specific reasoning — where accuracy, traceability, and auditability matter.

Your Responsibilities

• Data pipelines: Help build and improve ingestion and processing pipelines for insurance policy documents, legal texts, and case descriptions (parsing, normalization, structuring).
• AI‑ready structuring: Support chunking, metadata extraction, clause mapping, and schema design so downstream LLM agents work reliably against complex policy wording.
• LLM agent workflows: Prototype and implement agentic LLM workflows for coverage analysis — fact extraction, policy cross‑referencing, exclusion checking, and decision drafting — with attention to accuracy, cost, and latency.
• Search & retrieval: Assist with indexing strategies, embeddings, hybrid retrieval, and relevance tuning for legal and insurance content.
• Evaluation & quality: Build eval sets from real coverage check examples, run experiments, analyze failure cases, and help turn findings into shipped improvements.
• Engineering fundamentals: Write clean TypeScript code, add tests where appropriate, and improve observability (metrics/logging) for pipelines and services.
• Collaboration: Work closely with the founder and domain experts; document decisions and share learnings clearly.

Qualifications

• Near completion of your degree: You are at the end of your studies (e.g. finished Master's program or close to graduating).
• Degree in Computer Science, Data Science, Mathematics, or a related field (in progress) — or equivalent practical experience.
• Solid programming skills (preferably TypeScript/Node.js).
• Comfortable with data structures, algorithms, and working with real‑world data (edge cases, messy inputs).
• Working knowledge of SQL (ideally PostgreSQL).
• Interest in applied AI/ML systems (LLMs, embeddings, retrieval, evaluation, agentic workflows) and willingness to learn fast.
• Strong ownership mindset and clear communication in English.
• Availability full-time for the internship, ideally 6 months. On-site in Zurich at least twice per week.

Preferred qualifications

• Experience with one or more of: search systems (Azure AI Search / Elasticsearch), vector databases, or embeddings.
• Familiarity with data tooling (scraping/parsing, PDF tooling, OCR) and/or pipeline design.
• Experience with Docker and CI/CD basics.
• Familiarity with cloud platforms (especially Azure) is a plus.
• Working proficiency in German.
• Swiss work permit or EU/EFTA citizenship.

Benefits

• Direct impact: Your work ships into the product and directly improves how legal insurers process cases.
• Learning & mentorship: Hands‑on exposure to production LLM agent systems, search, and data pipelines in a high‑ownership environment.
• Breadth: Work across document ingestion → policy parsing → LLM reasoning → evaluation.
• Founder access: Work directly with the founder on architecture, product, and technical decisions.
• Compensation: CHF 2'500–4'500 per month, depending on experience and skills.
• Prospect of a permanent position: Possibility of full‑time employment with equity (ESOP) after a successful internship.