{"id":"CU25381","slug":"ai-driven-workflow-for-rapidly--CU25381","source":{"id":"CU25381","dataset":"techtransfer","title":"AI-driven workflow for rapidly generating custom operating systems","description_":"<p>This technology is a method for rapidly generating a custom operating system using a structured AI-driven workflow.</p>\r\r<h2>Unmet Need: Generating a custom operating system using AI</h2>\r\r<p>Custom operating systems are essential for a wide range of applications, including embedded devices, IoT platforms, and specialized hardware. However, developing these complex software systems is often slow, resource-intensive, and difficult to automate. While current AI-enabled coding assistants accelerate human task completion, they are limited to smaller, isolated tasks like function completion, and struggle with the scale and contextual demands of building an entire software system from the ground up. </p>\r\r<h2>The Technology: AI-driven workflow for rapidly generating custom operating systems</h2>\r\r<p>This technology describes a method called “Task Prompting,” which enables a human developer to effectively collaborate with a large language model (LLM) to generate a complete and functional operating system (OS) kernel. These prompts iteratively guide the LLM in creating all necessary C++ and assembly code for OS functionality on new hardware. This technology supports multi-process user application, virtual memory, systems calls, networking, and persistent storage. </p>\r\r<p>This technology was successfully used to generate a high-performance OS kernel that can run on unmodified Linux applications with performance comparable to native Linux.</p>\r\r<h2>Applications:</h2>\r\r<ul>\r<li>Rapid prototyping for IoT devices, robotics, and custom computer hardware</li>\r<li>Creation of custom operating systems for embedded systems</li>\r<li>AI-powered developer tools</li>\r<li>AI-tools for advanced educational platforms</li>\r<li>Enterprise software</li>\r<li>Computer engineering research model development</li>\r</ul>\r\r<h2>Advantages:</h2>\r\r<ul>\r<li>Rapid generation of working OS kernel</li>\r<li>Iterative process for creating, refining, and debugging codebase with large language models (LLMs)</li>\r<li>Supports multi-process user applications, virtual memory, system calls, networking, and persistent storage</li>\r<li>Utilizes structured methodology for human-AI interactions</li>\r</ul>\r\r<h2>Lead Inventor:</h2>\r\r<p><a href=\"https://www.engineering.columbia.edu/faculty-staff/directory/jason-nieh\">Jason Nieh, Ph.D.</a></p>\r\r<h2>Patent Information:</h2>\r\r<p>Patent Pending</p>\r\r<h2>Related Publications:</h2>\r\r<h2>Tech Ventures Reference:</h2>\r\r<ul>\r<li><p>IR CU25381</p></li>\r<li><p>Licensing Contact: <a href=\"mailto:techtransfer@columbia.edu\">Greg Maskel</a> </p></li>\r</ul>\r","tags":["Artificial intelligence","Assembly language","Computer engineering","Computer hardware","Embedded system","Enterprise software","Internet of things","Language model","Linux kernel","Operating system","Rapid prototyping","Virtual memory","Workflow"],"file_number":"CU25381","collections":[],"meta_description":"AI-driven workflow rapidly generates custom OS kernels via structured prompts, enabling fast, multi-process, networked, persistent systems.","apriori_judge_output":"{\"scores\":{\"novelty\":4.0,\"potential_impact\":5.0,\"readiness\":4.0,\"scalability\":4.0,\"timeliness\":4.0},\"weighted_score\":4.2,\"risks\":[\"High reliance on LLMs may introduce security and correctness risks\",\"Potential for patentability challenges due to broad AI-assisted code generation\",\"Competition from established OS development tooling\",\"Ethical/dual-use considerations for OS-level modifications\"],\"one_sentence_take\":\"Strong novelty and impact with near-term readiness, but faces AI safety, security, and integration risks that should be mitigated before scaling.\"}","inventors":["Jason Nieh"],"manager":"Greg Maskel","depts":["Computer Science"],"divs":["Fu Foundation School of Engineering and Applied Science (SEAS)"],"date_released":"2026-07-24"},"highlight":{},"matched_queries":null,"score":0.0}