Spare Parts 3D – Summer Internship – AI & Data Engineering Intern – Operational Intelligence

Spare Parts 3D, headquartered in France with offices in Singapore and Qatar, is revolutionizing the spare parts industry through cutting-edge 3D printing solutions. Serving sectors like Railway, Energy, and Semiconductors, we enable the digitization of spare parts inventories via our proprietary DigiPART* software and AI-powered visual engine Theia. Our data-driven approach informs strategic decisions on Additive Manufacturing (AM), with global on-demand production capabilities across Australia, Europe, Asia, and the Americas — reducing warehousing costs and streamlining supply chains.

If you're ready to take on real responsibilities, thrive in fast-paced environments, and want to help build the next generation of supply chain technology, this is your opportunity to join a high-impact team!

Overall Objective

Explore and prototype the use of AI to enhance internal operational pipelines and customer-facing deliverables, while gaining exposure to real-world data and domain-specific challenges.

Workstream Overview

1. Pipeline Familiarization & Model Benchmarking (Weeks 1–3)

Goal: Understand the current operational AI pipeline and establish a performance baseline.

Deliverables:

• Review internal workflows and models used (e.g., classification, matching, extraction).
• Evaluate existing outputs using confusion matrix and real-world examples.
• Present a short report highlighting performance, gaps, and improvement priorities.

2. AI-Powered Dashboard & Report Assistant (Weeks 4–8)

Goal: Design and prototype an AI assistant that generates summaries, alerts, and standardized reports for customers.

Deliverables:

• Analyze current reporting workflows and formulate standardized report structures.
• Build a basic AI agent to generate insights or automate reporting using tools such as code interpreters, web search/scraping, or LLMs.
• Integrate a lightweight interface (e.g., Streamlit).

• Conduct 1–2 stakeholder demos and incorporate feedback.

3. Visual BOM Data Extraction – Use Case Deep Dive (Weeks 6–10)

Goal: Begin tackling the complex problem of extracting structured Bill of Material (BOM) data from visual inputs such as technical drawings or PDF.

Deliverables:

• Focus on 1–2 use cases (e.g., CAD exports, scanned drawings).
• Build a prototype using OCR, layout parsing, or visual AI models.

• Measure accuracy, identify limitations, and document findings.

Tech Exposure

• Languages: Python, SQL
• Libraries: HuggingFace, OpenCV, LangChain, Tesseract, Pandas, Polars
• Tools: Streamlit, GitLab, Postgres, Microsoft Power BI, Microsoft Office
• AI Services (Optional): OpenAI API, UiPath, Gemini AI , Docker, Google Cloud , AWS 

Expected Achievements

• Familiarity with internal AI pipelines and operational metrics.
• MVP of a reporting assistant with real stakeholder feedback.
• Prototype for extracting structured BOM data from at least one visual format.

 Ideal Candidate

• Final-year or graduate-level student in AI, Data Science, or Software Engineering.
• Practical experience with Python and basic machine learning.
• Eagerness to tackle real-world, messy data challenges.
• Bonus: Interest in manufacturing, visual data, or operational processes.