How I built my own intelligent Robot Arm from Scratch - Iulia Feroli - NDC Oslo 2026

How I built my own intelligent Robot Arm from Scratch - Iulia Feroli - NDC Oslo 2026

Title history
  • How I built my own intelligent Robot Arm from Scratch - Iulia Feroli - NDC Oslo 2026 — observed on 20261025-1532 UTC (YouTube)
🎙 Iulia Feroli 👥 230K 📅 October 7, 2026 ⏱ 38 min 👁 2K 📄 tutorial 🧭 2026-10-25
Available in: English (current) Français

Keywords

physical AIrobot armteleoperation3D printingservo motors

Summary

Iulia Feroli, a former data scientist, shares her journey of building a robotic arm from scratch, driven by a desire to apply AI to physical systems. She begins with low-cost components like Arduino and Raspberry Pi microcontrollers, inexpensive servo motors, and scrap materials like cardboard, gradually progressing to 3D-printed parts. She emphasizes the educational value of building without buying expensive kits, and demonstrates how to create a functional robot arm for about $20. The talk covers basic electronics, soldering, and code control, and then introduces the concept of ‘physical AI’ as coined by Nvidia, integrating LLMs into robotics. She showcases teleoperation using a Hugging Face ‘Le Robot’ design, where two arms mirror each other, and explains how recorded movements can become training data for autonomous behavior. The session includes live demonstrations of recording and playing back a movement, and concludes with thoughts on the future of accessible robotics and AI.

150 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk offers substantial practical value for hobbyists and beginners interested in robotics and AI. The speaker’s step-by-step approach demystifies the process, highlighting affordable alternatives and common pitfalls. Her argumentation is grounded in personal experience, which lends authenticity but also limits generalizability. She honestly discusses failures, such as twitchy movements and unstable cables, reinforcing the credibility of her methods. The case for combining LLMs with robotics is persuasive, though the technical depth is moderate, aimed at a general technical audience.

Scientific Rigor, Source Quality, Title Accuracy

The talk references Nvidia as the origin of ‘physical AI’ and Hugging Face’s Le Robot open-source design, but provides no formal citations. The sources mentioned are credible and well-known within the tech community. The title matches the content precisely, as the speaker indeed builds an intelligent robot arm from scratch, culminating in a teleoperation demo and movement recording. The scientific rigor is moderate: the talk is an anecdotal account rather than a controlled study, but it aligns with established practices in hobbyist robotics and AI. The description includes only conference links, which do not serve as substantive sources.

193 words

Title / Content Match

The title accurately reflects the content: the speaker details her process of building a robot arm from scratch and eventually adding AI capabilities.

Quality & Reliability

7/10

The talk provides a practical, hands-on account of building a robot arm, with clear explanations and real demonstrations. The speaker is experienced and shares both successes and failures, making the content credible. However, it is a personal narrative rather than a rigorous scientific study, and some technical details are simplified.

Key Moments

Cited Sources

  • NDC Conferences — Main conference website where the talk was recorded.
  • NDC Oslo — Specific event page for the conference.

Contribution & Novelties

The talk provides an accessible, low-cost pathway for building a robot arm and integrating AI, bridging the gap between software and hardware. The emphasis on teleoperation and data collection for future autonomous models offers a fresh perspective.

Pour aller plus loin :

94 words

Radar Profile

The profile is balanced across dimensions, with slightly lower scores in technical depth due to the simplified explanations. The talk excels in practical applicability and community engagement, making it a strong entry point for beginners, while lacking the rigor of formal research.

Reliability 7/10