Promoting FLOSS through FOSSEE and Spoken Tutorial (Documentation, Community & Open Source)

• Kannan Mougalya (IIT Bombay)

Room: Room 112 Digital Sovereignty has become a critical requirement in these days of tariffs and sanctions. Free/Libre and Open Source Software is a tool to achieve this freedom. FOSSEE and Spoken Tutorial were started at IIT Bombay by the Ministry of Education to promote FLOSS. This activity has resulted in training of ten million students and teachers over a period of 12 years. The training method developed by Spoken Tutorial has received an IEEE Standard. This is the only educational technology from India to receive an IEEE Standard.

eSim on Ubuntu: Open Source EDA and PDKs for Circuit Design with Snap, Docker, and AI (Embedded Systems & IoT)

• Sumanto Kar (FOSSEE, IIT Bombay)
• Varad Patil (FOSSEE, IIT Bombay)
• Shanthi Priya (FOSSEE, IIT Bombay)
• Aditya Minocha (FOSSEE, IIT Bombay)

Room: Room 112 [eSim][1] on Ubuntu represents a complete open-source ecosystem for electronic design, simulation, and prototyping, combining the flexibility of Ubuntu with the power of modern packaging and deployment tools like Snap and Docker. Developed at [FOSSEE][2], IIT Bombay, eSim integrates schematic capture, circuit simulation using Ngspice and Verilator, mixed-signal support, and PCB workflows through KiCad, providing a unified alternative to costly proprietary EDA tools. With Snap packaging, users benefit from simple installation and seamless updates across Ubuntu versions, while Docker containers ensure reproducibility, portability, and ease of setup for workshops, collaborative research, and cloud environments. The inclusion of open-source Process Design Kits (PDKs) such as SkyWater 130nm, GF180(Work in Progress), and IHP SG13G2 bridges the gap between education and industry by allowing users to take their designs from circuit level to silicon, fostering innovation in open hardware. Looking ahead, the convergence of AI with eSim on Ubuntu opens new possibilities for automated circuit optimization, intelligent design space exploration, and smart error detection, making the design process faster, more efficient, and more accessible. Together, eSim, open-source PDKs, and Ubuntu’s AI-ready ecosystem pave the way for a democratized future in electronics design—one where students, researchers, startups, and industries can all collaborate and innovate without barriers. All the code resource for eSim are available on [eSim-GitHub][3] page. We are developing a [Cloud version of eSim][4]. [1]: https://esim.fossee.in [2]: https://fossee.in [3]: https://github.com/FOSSEE/eSim [4]: https://github.com/frg-fossee/eSim-Cloud

Building Production-Ready Kubernetes on Ubuntu: A Checklist Approach (DevOps & Security)

• Kunal Das (CastAI)

Room: Room 112 Moving from development to production Kubernetes on Ubuntu involves critical decisions that can make or break your deployment. This practical session provides an actionable checklist transforming production readiness into manageable steps. Drawing from real-world deployments, we'll cover essential components of production-ready Ubuntu Kubernetes clusters: CNI selection, storage strategies, CIS-aligned security hardening, and monitoring architectures that provide visibility without complexity. Attendees receive a downloadable checklist covering networking, storage classes, RBAC implementation, observability stack deployment, disaster recovery, and automated security scanning. Each item includes recommended open-source tools, Ubuntu-specific configurations, and common pitfalls. We'll demonstrate eBPF-based runtime monitoring, automated vulnerability scanning, cost-effective logging, and CI/CD pipelines with security policies. The session covers leveraging Ubuntu's AppArmor in containers, resource optimization, and scalable monitoring from small clusters to multi-region deployments. Focus is on actionable insights you can implement immediately: kubectl commands, YAML configurations, and automation scripts for Ubuntu environments. Help attendees avoid common mistakes and build robust, secure, maintainable platforms.

Ubuntu on Qualcomm Dragonwing platform for EdgeAI applications

• Sujai SyrilRaj (Qualcomm)

Room: Room 112 In this session, we introduce two developer boards built on the Qualcomm Dragonwing platform, both supporting Ubuntu 24.04: Rubik Pi 3: Showcases AI/ML workflows with Edge Impulse, optimized for real-time inference and seamless sensor integration. IQ9 SBC: Delivers high-performance capabilities for multimedia applications, industrial control, and advanced edge analytics.

Edge AI applications on Ubuntu, powered by the Qualcomm Dragonwing platform

• Ravi Kumar Neti (Qualcomm)

Room: Room 112 This session explores real-world Edge AI applications using Qualcomm developer boards. Key use cases include: Deploying ML models for predictive maintenance and anomaly detection. Enabling smart vision and voice interfaces on the IQ9 SBC.

Closing Remarks

• Bhavanishankar Ravindra (Private Entrepreneur/Ubuntu developer)

Don't just write documentation, design it! (Documentation, Community & Open Source)

• Keirthana Thangavelu Srinivasan (Canonical)

Room: Room 112 This talk will focus on benefits of designing documentation based on the user experience. What typically happens in the world of documentation, especially in open source communities is that documentation is written "for" a product, keeping the product evolution front and center. This talk intends to challenge that notion by comparing real time examples and demonstrating how consciously designing documentation following the user journey instead of the product journey can provide better positioning for the products in the market and also draw users' attention better.

Why Young India Needs To Rock Open Source, Now. (Cloud & Infrastructure)

• Soundarya Rangarajan (Canonical)

Room: Room 112 India’s developer ecosystem is one of the youngest and fastest-growing in the world, yet there’s still a gap between brilliant blue-sky ideas and truly enterprise-grade, sustainable solutions. At the same time, vendor lock-in, rising cloud costs, and security concerns make it critical for India to take ownership of its digital future. Most young developers here already work with Ubuntu, making it one of the most popular Linux base images for containers. But what if there was a better way to go beyond just building apps, towards building secure, scalable solutions that stand the test of time? This talk introduces **Rocks**—Ubuntu’s next-gen container images—and shows why they matter for India. We’ll explore: 1. How Rocks differ from familiar Docker images. 2. How **Chisel** tooling reduces attack surfaces without compromising developer experience. 3. How Rocks tie into the **broader Ubuntu ecosystem** (Juju, charms, orchestration) to unlock scalable, multi-cloud, vendor-neutral deployments—the very model already powering national projects like UIDAI’s microservices platform for Aadhar. 4. A **demo** of an **India-relevant use case involving DPGs (Digital Public Goods)**, showing how simple it is to start “rocking” your containerized apps, and what implications this has on security, stability and lifecycle-management in the long run. By the end, you’ll see how Rocks can be the foundation for India’s young innovators to build open source that’s not only developer-friendly, but also secure, sovereign, and enterprise-grade.

Developer-Ready Ubuntu on Qualcomm IoT Platforms

• LK Reddy Ponnatota (Qualcomm)

Room: Room 112 Discover how Ubuntu and Linux open-source solutions power Qualcomm SoCs and hardware kits such as Rubik Pi, Arduino Uno Q, and IQ9 EVK. This session will showcase software offerings for AI/ML development, including demos of AI vision models and small language models running on Ubuntu on-device. Learn how these platforms enable building innovative edge AI use cases across diverse markets—retail, smart homes, industrial automation, and more—using IoT devices at the edge. LK Reddy Ponnatota is Senior Director of Engineering at Qualcomm, leading initiatives in Edge AI solutions across business units. With extensive experience in AI/ML enablement on Qualcomm platforms, LK has driven developer-focused innovations across AI/ML workflows, GenAI, applications, and edge deployments. LK has spearheaded projects integrating Ubuntu with Qualcomm SoCs like Rubik Pi, Arduino Uno Q, and IQ9 EVK, enabling developers to build AI vision models and small language models on-device. LK’s work empowers industries to adopt open-source solutions for edge AI use cases in retail, smart homes, and industrial automation. At UbuCon India, LK will present “Developer-Ready Ubuntu on Qualcomm IoT Platforms”, showcasing AI workflows, demos, and best practices for building edge intelligence using Ubuntu.

OSCAL - An free open source DSL for GRC Engineers (DevOps & Security)

• Balachandran Sivakumar

Room: Room 112 OSCAL (Open Security Controls Assessment Language), developed by NIST, is rapidly becoming the standard for representing compliance, controls, risks, and assessment data in machine-readable formats. While already gaining traction in FedRAMP and other GRC automation efforts, its potential for the broader cybersecurity ecosystem is only beginning to unfold. In this talk, I will share extensions to OSCAL that enable more efficient handling of multi-framework conformance and crosswalks between standards. I will also show how OSCAL can be applied to audit reporting, making reports not just human-readable but also machine-parseable and automation-friendly. These advancements open the door to faster compliance workflows, reduced manual effort, and a more open, interoperable approach to GRC.

Data Center Virtualization 2025 (Cloud & Infrastructure)

• Saquib Akhtar (Linux SME & IDM L3 Lead @ NOKIA)

Room: Room 112 Virtualization is at the heart of today’s data centers, and Ubuntu plays a key role in making it accessible and efficient. From KVM and QEMU to cloud platforms and containerization, Ubuntu provides the foundation for building flexible, scalable, and cost-effective virtualized environments. In this session, I will share experiences and practical use cases of deploying and managing virtualization on Ubuntu, focusing on how open-source tools can simplify operations, improve resource utilization, and empower developers and sysadmins alike.

Open Source Chip Design Ecosystem (Documentation, Community & Open Source)

• N Nishchit (BMS College of Enigneering)

Room: Room 112 The open-source chip design ecosystem is one of the way everyone can learn how hardware is designed, verified, and fabricated. This talk will provide an introductory overview of key open-source tools, frameworks, and processes, focusing on the *RISC-V architecture*, *SkyWater PDK*, and the *OpenROAD autonomous design flow* and some other supporting tools. Through my personal exploration, I will highlight how beginners can start contributing, experimenting, and learning in this ecosystem. **Attendees will leave with a clear roadmap of resources, practical examples, and a foundational understanding of open-source chip design.**

GitOps + Ubuntu: Managing Infrastructure, Snap & Cloud Deployments Declaratively (Cloud & Infrastructure)

• Neel Shah

Room: Room 105 GitOps is not limited to Kubernetes - it is a paradigm for managing everything as code. In this dialogue, we will find out how GitOps practices with Flux or ArgoCD can be extended to the Ubertu system itself. From OS configuration and security patch to Snap, APT package and management of service, GIT becomes a single source of truth. We will showcase how to control the Ubuntu infrastructure on the cloud, on-dim or edge, and show how automatic rollouts, drift detection and rollback create a self-healing, auto-environment. The attendees will leave the app and infrastructure management under the GitOps with a clear understanding of the way to unite, bridging the gap between the cloud native workflow and the wide Ubuntu ecosystem.

Qualcomm IoT Platforms: Journey from Yocto to Ubuntu

• Sourabh Banerjee (Qualcomm)

Room: Room 112 An overview of what Qualcomm engineering teams have done to provide Ubuntu based development experience on Qualcomm IoT Platforms. Sourabh Banerjee has been part of Qualcomm for over 15 years, playing a part in shaping the future of embedded systems and smart devices. A seasoned engineering leader, Sourabh has led and contributed to groundbreaking projects — including one of the first smartwatches powered by Snapdragon. With over a decade of experience working on Linux-based distributions, Sourabh has been instrumental in enabling Qualcomm SoCs to power a wide range of products, from IP and surveillance cameras to voice assistants and other connected devices. His work continues to be in the area of bringing intelligent edge computing to life.

The Heartbeat of Open Source: Stories, Pathways and People Beyond Code (Documentation, Community & Open Source)

• Aditya D
• Saumili Dutta

Room: Room 112 When people talk about open source, they usually think it is only commits, pull requests, and code brilliance. While the truth is, code isn’t the whole story. What really keeps open source alive are the ecosystems of care, culture, and community wrapped around it. Behind every healthy project, there’s a README that introduces the project, Contribution guidelines that invites them in, Communities—local or global—that meets, helps and grows together. Documentation is no different than storytelling where Licensing reflects values and philosophy and diversity and inclusion become the actual foundation of resilience. This talk challenges the idea that contribution = code. Drawing from real experiences in FOSS projects and community building, we’ll share how we can: Write documentation that feels like an invitation, not an instruction manual. Build and sustain local communities (LoCos) that strengthen advocacy and amplify diverse voices. Design contribution pathways that act like bridges, not walls. Breathe diversity and inclusion into the DNA of projects, not tack them on as extras. Treat licensing and governance as tools that empower, not intimidate. Learning Outcomes Recognize non-code contributions as critical to FOSS sustainability. Learn how inclusive documentation and guidelines lower barriers. Understand strategies for building and supporting diverse local communities. See licensing and governance as cultural as well as legal choices. Leave with practical steps to nurture open source as an ecosystem of belonging

Low-Latency OpenAFS on Ubuntu: Real-Time File Sharing for Robots & Edge Devices (Embedded Systems & IoT)

• Utkarsh Maurya (National Institute of Technology, Hamirpur, Himachal Pradesh, India)

Room: Room 112 **Low-Latency OpenAFS on Ubuntu: Real-Time File Sharing for Robots & Edge Devices** **Abstract** Here’s the deal: robots and sensors are everywhere now. Drones, Raspberry Pis, Jetsons, Coral boards - you name it. They’re all spitting out massive amounts of data: LiDAR scans, video feeds, GPR readings. The problem? Getting that data across multiple compute nodes fast, reliably, and without crying about cloud latency. That’s where OpenAFS comes in. Yeah, it’s this old-school distributed file system that most people forgot about, but when you hack it right, it’s insanely powerful. And when you throw Ubuntu into the mix (on both ARM and x86), you get this battle-tested, offline-friendly, low-latency pipeline for robotics and embedded systems. In this talk, I’ll show you how I’ve been using OpenAFS to do things like: - Stream LiDAR data from drones in real time without choking. - Collect and sync sensor logs across a fleet of Raspberry Pis. - Tie it all into ROS so multiple agents share data like they’ve got a group chat. - Tune the Ubuntu kernel and network stack so the system doesn’t jitter when milliseconds actually matter. And yes - we’ll go through real field-tested use cases, like using aerial drones with Ground Penetrating Radar for landslide rescue. It’s not theory. It’s stuff I’ve actually hacked on, crashed on, and debugged at 3AM. If you care about robotics, real-time systems, or just want to see how an “ancient” distributed file system can unlock futuristic applications, you’ll want to catch this. **What You’ll Take Away** 1. How to deploy and configure OpenAFS on Ubuntu for edge environments. 2. Tricks to shave latency off file read/writes on low-power devices. 3. How distributed file systems can simplify robotics and IoT data flows. 4. Why you don’t always need a cloud pipeline to do serious real-time stuff. **Who This Talk Is For** Intermediate Linux nerds, robotics engineers, system builders, or anyone who’s ever had to move a ton of sensor data between edge devices and thought: “Why is this so painful?” **Prereqs** Know your way around Ubuntu. Basic networking/file systems knowledge. If you’ve touched a Raspberry Pi or messed with ROS, even better. **Speaker Bio** I’m **Utkarsh Maurya** - senior at NIT Hamirpur, ex-firmware dev at Cypherock, and currently a kernel dev at a stealth startup. I’ve built drone-based landslide rescue systems with custom GPR, done low-latency sensor networking on Raspberry Pi/Jetson/Coral boards. I also mess around with bare-metal 6502 computers, blockchain infra, and kernel internals for fun. Basically, I like taking weird open-source tech and pushing it way past what people think it’s for.

Streamlining Pedagogical Content Creation- A Hierarchical Agent-Based Architecture for the Procedural Generation of Multimodal FOSS Tutorials (Artificial Intelligence & Machine Learning (AI/ML))

• Sumanto Kar (FOSSEE, IIT Bombay)
• Aditya Bhattacharya (Rajiv Gandhi Institute of Petroleum Technology)
• Kannan Moudgalya (Director EduPyramids, Former Professor IIT Bombay)

Room: Room 105 We introduce a cognitive architecture for the autonomous synthesis of pedagogical content, inspired by the [Spoken Tutorial project][1] and aligned with the broader goals of the [FOSSEE][2] (Free/Libre and Open Source Software for Education) initiative at IIT Bombay. This framework employs a distributed agentic workflow to automate the end-to-end generation of instructional materials for Free and Open Source Software (FOSS). The primary objective is to address the significant latency, resource overhead, and manual fidelity required by conventional content creation methodologies. Our system is engineered to radically compress the production lifecycle, thereby scaling the availability of high-quality FOSS education for academic and resource-constrained environments. Conventional didactic content production is characterized by a disarticulated, high-latency workflow that is heavily reliant on human expertise at multiple stages. This paradigm presents substantial barriers to scalability and introduces variability in pedagogical quality. In contrast, our proposed architecture operationalizes the entire creation process as a unified, deterministic computational problem. It transforms a high-level topic directive into a complete, synchronized suite of instructional assets, thereby enforcing a uniform stylistic and pedagogical standard while minimizing human intervention. The system is architected as a logical pipeline implemented across a distributed workflow of functionally specialized agents. Task execution is coordinated asynchronously through a unified communication channel, which governs the sequential progression of artifact generation. The process is initiated when a Pedagogical Scripting Agent consumes an initial topic directive and produces a complete narrative script. This artifact is then published to the channel, triggering a human-in-the-loop (HITL) validation gate where the script undergoes review and potential refinement. Upon successful validation, the verified script is broadcast on the channel, where it is consumed by the Visual Manifestation Agent. This agent then generates a semantically aligned set of presentation slides. In the terminal stage of the pipeline, both the script and its corresponding slides are consumed by a Temporal Synchronization Agent, which computes and fuses the final timing metadata to produce the complete, synchronized tutorial package. The realization of this distributed architecture was pivotally dependent on the Ubuntu environment. Its native support for kernel-level containerization primitives and high-throughput networking APIs provided the bedrock for our asynchronous communication channel, while its role as the reference platform for leading AI frameworks ensured seamless integration of the entire development toolchain. The cognitive core of each agent is powered by a sophisticated Retrieval-Augmented Generation (RAG) paradigm. To ensure maximal contextual relevance, we implement a hybrid retrieval strategy, combining sparse (e.g., BM25) and dense vector search mechanisms to query a domain-specific corpus of FOSS documentation, existing tutorials, and visual design templates curated under the FOSSEE initiative. The retrieved document candidates and graphical precedents are subsequently re-ranked by a lightweight cross-encoder model before being injected into the prompt. For complex problem-solving, agents leverage a Chain-of-Thought (CoT) reasoning process. This enables them to generate intermediate logical steps before producing the final output, critical for structuring coherent narratives and designing semantically relevant visual progressions. Furthermore, a Chain-of-Verification (CoVe) process is integrated to iteratively refine generated content against retrieved facts and pre-defined pedagogical criteria, ensuring both factual grounding and logical, pedagogical soundness in the synthesized artifacts. By synergizing a collaborative multi-agent topology with an advanced hybrid-retrieval RAG framework, Chain-of-Thought reasoning, and a visual synthesis component, our system provides a novel solution for autonomous instructional design. The architecture drastically reduces the content creation lifecycle, enhances pedagogical uniformity, and democratizes the capacity to produce high-quality FOSS tutorials at scale. This work builds upon and extends the long-standing contributions of FOSSEE, IIT Bombay, in creating accessible, open-source educational resources, offering a scalable model for academic and open-source communities worldwide. Keywords: AI, Chain-of-Verification, RAG, Multi-Agent Systems, Chain-of-Thought (CoT), LLM, Automation, Educational Technology, FOSS, Ubuntu, HITL, Multimodal Generation. [1]: https://spoken-tutorial.org [2]: https://fossee.in

Introducing Canonical Academy (Documentation, Community & Open Source)

• Ishani Ghoshal

Room: Room 112 In this talk, we want to introduce [Canonical Academy][1] - a new learning and qualification platform built to empower individuals with real-world, hands-on skills that reflect the way open source is used in production today. We’ll give the community an inside look at the development of Canonical Academy, how it aligns with Ubuntu’s mission, and why we believe it fills an important gap in the current Linux and open source learning ecosystem. What the session will cover: 1. How Canonical Academy is structured - An overview of our job-based tracks, starting with the System Administrator qualification, and what learners can expect from each. 2. Our exam creation methodology - A behind-the-scenes look at how we design exams to be practical, meaningful, and reflective of the actual work done in professional environments — including input from engineers, educators, and subject matter experts. 3. What this means for the community - how to validate skills, upskill for a job, contribute as an SME. 4. Get involved- Learn how to join as a beta tester, sign up as a subject matter expert, provide feedback to shape future exam tracks [1]: https://ubuntu.com/credentials

Embedded UAV Robotics in the Ubuntu Ecosystem: Robotic Arm Integration for Smart Delivery Applications (Embedded Systems & IoT)

• Karthik M

Room: Room 112 ### Abstract The final frontier of logistics,last-mile delivery is ripe for disruption. This presentation introduces an advanced smart delivery UAV with an integrated robotic arm, a system built from the ground up on the foundational tools of the **Ubuntu ecosystem**. We demonstrate how Ubuntu provides the stability, security, and flexibility required for complex, real-world robotics applications. We will explore the complete project lifecycle, showcasing why **Ubuntu** is the platform of choice for serious robotics development. From the long-term support of **Ubuntu 22.04 LTS** during the build phase to the transactional updates of **Ubuntu Core** in deployment, the system's architecture pivots on the deep integration between Ubuntu and the Robot Operating System (**ROS 2**). This session will detail the architecture, covering: - **An Ubuntu-Centric Integration:** How **Ubuntu** provides the stable and performant environment for **ROS 2** to seamlessly integrate the flight controller with the robotic arm's servos. - **Validated Mechanical Design:** A look into the engineering analysis, including Finite Element Analysis (FEA) on the carbon fiber frame and the numerical calculations that determined a stable 2:1 thrust-to-weight ratio. - **Mission Control on Ubuntu:** Using the popular Mission Planner software on an **Ubuntu-based ground station** to program complex, autonomous delivery sequences using MAVLink commands. - **Deployment with Ubuntu Core & Snaps:** Why **Ubuntu Core** is the ideal OS for reliable edge devices, using **Snaps** to containerize the drone's software for enhanced security and robust, over-the-air updates. Attendees will leave with a clear, practical understanding of why **Ubuntu** has become the cornerstone of the robotics industry. This talk is for developers, engineers, and hobbyists who want to leverage **Ubuntu, Ubuntu Core, and ROS 2** to build their own sophisticated autonomous systems.

Closing Remarks

• Bhavanishankar Ravindra (Private Entrepreneur/Ubuntu developer)

Room: Room 112